<?xml version="1.0"?>
<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="en">
	<id>https://onnocenter.or.id/wiki/index.php?action=history&amp;feed=atom&amp;title=Keras%3A_Python_Keras_Text_Classification</id>
	<title>Keras: Python Keras Text Classification - Revision history</title>
	<link rel="self" type="application/atom+xml" href="https://onnocenter.or.id/wiki/index.php?action=history&amp;feed=atom&amp;title=Keras%3A_Python_Keras_Text_Classification"/>
	<link rel="alternate" type="text/html" href="https://onnocenter.or.id/wiki/index.php?title=Keras:_Python_Keras_Text_Classification&amp;action=history"/>
	<updated>2026-04-29T18:30:45Z</updated>
	<subtitle>Revision history for this page on the wiki</subtitle>
	<generator>MediaWiki 1.35.4</generator>
	<entry>
		<id>https://onnocenter.or.id/wiki/index.php?title=Keras:_Python_Keras_Text_Classification&amp;diff=56758&amp;oldid=prev</id>
		<title>Onnowpurbo: /* Penggunaan Pretrained Word Embeddings */</title>
		<link rel="alternate" type="text/html" href="https://onnocenter.or.id/wiki/index.php?title=Keras:_Python_Keras_Text_Classification&amp;diff=56758&amp;oldid=prev"/>
		<updated>2019-08-14T01:30:32Z</updated>

		<summary type="html">&lt;p&gt;&lt;span dir=&quot;auto&quot;&gt;&lt;span class=&quot;autocomment&quot;&gt;Penggunaan Pretrained Word Embeddings&lt;/span&gt;&lt;/span&gt;&lt;/p&gt;
&lt;table class=&quot;diff diff-contentalign-left diff-editfont-monospace&quot; data-mw=&quot;interface&quot;&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;tr class=&quot;diff-title&quot; lang=&quot;en&quot;&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;← Older revision&lt;/td&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;Revision as of 01:30, 14 August 2019&lt;/td&gt;
				&lt;/tr&gt;&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l634&quot; &gt;Line 634:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 634:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;                  idx = word_index[word]  &lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;                  idx = word_index[word]  &lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;                  embedding_matrix[idx] = np.array(&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;                  embedding_matrix[idx] = np.array(&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;                     &lt;/del&gt;vector, dtype=np.float32)[:embedding_dim]&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;                         &lt;/ins&gt;vector, dtype=np.float32)[:embedding_dim]&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt; &lt;/del&gt;&lt;/div&gt;&lt;/td&gt;&lt;td colspan=&quot;2&quot;&gt; &lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;      return embedding_matrix&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;      return embedding_matrix&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l642&quot; &gt;Line 642:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 641:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  embedding_dim = 50&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  embedding_dim = 50&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  embedding_matrix = create_embedding_matrix(&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  embedding_matrix = create_embedding_matrix(&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;     &lt;/del&gt;'&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;data&lt;/del&gt;/&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;glove_word_embeddings&lt;/del&gt;/glove.6B.50d.txt',&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;         &lt;/ins&gt;'/&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;home/onno/TensorFlow/glove.6B&lt;/ins&gt;/glove.6B.50d.txt',&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;     &lt;/del&gt;tokenizer.word_index, embedding_dim)&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;         &lt;/ins&gt;tokenizer.word_index, embedding_dim)&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;Sekarang kita siap menggunakan embedding matrix dalam training. Mari gunakan network sebelumnya dengan global max pooling dan lihat apakah kita bisa meningkatkan model ini. Saat kita menggunakan pretrained word embeddings  kita memiliki pilihan untuk memperbolehkan embedding diperbarui selama training atau hanya menggunakan embedding vectors yang dihasilkan apa adanya.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;Sekarang kita siap menggunakan embedding matrix dalam training. Mari gunakan network sebelumnya dengan global max pooling dan lihat apakah kita bisa meningkatkan model ini. Saat kita menggunakan pretrained word embeddings  kita memiliki pilihan untuk memperbolehkan embedding diperbarui selama training atau hanya menggunakan embedding vectors yang dihasilkan apa adanya.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l650&quot; &gt;Line 650:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 649:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  nonzero_elements = np.count_nonzero(np.count_nonzero(embedding_matrix, axis=1))&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  nonzero_elements = np.count_nonzero(np.count_nonzero(embedding_matrix, axis=1))&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  nonzero_elements / vocab_size&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;print(&lt;/ins&gt;nonzero_elements / vocab_size&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;)&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;0.&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;9507727532913566&lt;/del&gt;&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt; &lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot;&gt; &lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Hasilnya:&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot;&gt; &lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt; &lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot;&gt; &lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt; &lt;/ins&gt;0.&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;9522330097087378&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;Ini berarti 95.1% dari vocabulary tercakup oleh pretrained model, yang merupakan cakupan yang baik dari vocabulary kita. Mari kita lihat kinerja ketika menggunakan lapisan GlobalMaxPool1D:&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;Ini berarti 95.1% dari vocabulary tercakup oleh pretrained model, yang merupakan cakupan yang baik dari vocabulary kita. Mari kita lihat kinerja ketika menggunakan lapisan GlobalMaxPool1D:&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;/table&gt;</summary>
		<author><name>Onnowpurbo</name></author>
	</entry>
	<entry>
		<id>https://onnocenter.or.id/wiki/index.php?title=Keras:_Python_Keras_Text_Classification&amp;diff=56757&amp;oldid=prev</id>
		<title>Onnowpurbo: /* Penggunaan Pretrained Word Embeddings */</title>
		<link rel="alternate" type="text/html" href="https://onnocenter.or.id/wiki/index.php?title=Keras:_Python_Keras_Text_Classification&amp;diff=56757&amp;oldid=prev"/>
		<updated>2019-08-14T01:25:05Z</updated>

		<summary type="html">&lt;p&gt;&lt;span dir=&quot;auto&quot;&gt;&lt;span class=&quot;autocomment&quot;&gt;Penggunaan Pretrained Word Embeddings&lt;/span&gt;&lt;/span&gt;&lt;/p&gt;
&lt;table class=&quot;diff diff-contentalign-left diff-editfont-monospace&quot; data-mw=&quot;interface&quot;&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;tr class=&quot;diff-title&quot; lang=&quot;en&quot;&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;← Older revision&lt;/td&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;Revision as of 01:25, 14 August 2019&lt;/td&gt;
				&lt;/tr&gt;&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l617&quot; &gt;Line 617:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 617:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;Ini adalah file besar dengan 400.000 baris, dengan setiap baris mewakili kata yang diikuti oleh vektornya sebagai aliran floating point. Misalnya, berikut adalah 50 karakter pertama dari baris pertama:&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;Ini adalah file besar dengan 400.000 baris, dengan setiap baris mewakili kata yang diikuti oleh vektornya sebagai aliran floating point. Misalnya, berikut adalah 50 karakter pertama dari baris pertama:&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  $ head -n 1 &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;data/glove_word_embeddings&lt;/del&gt;/glove.6B.50d.txt | cut -c-50&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  $ head -n 1 &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;glove.6B&lt;/ins&gt;/glove.6B.50d.txt | cut -c-50&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;      the 0.418 0.24968 -0.41242 0.1217 0.34527 -0.04445&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;      the 0.418 0.24968 -0.41242 0.1217 0.34527 -0.04445&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;/table&gt;</summary>
		<author><name>Onnowpurbo</name></author>
	</entry>
	<entry>
		<id>https://onnocenter.or.id/wiki/index.php?title=Keras:_Python_Keras_Text_Classification&amp;diff=56753&amp;oldid=prev</id>
		<title>Onnowpurbo: /* Pranala Menarik */</title>
		<link rel="alternate" type="text/html" href="https://onnocenter.or.id/wiki/index.php?title=Keras:_Python_Keras_Text_Classification&amp;diff=56753&amp;oldid=prev"/>
		<updated>2019-08-14T01:01:06Z</updated>

		<summary type="html">&lt;p&gt;&lt;span dir=&quot;auto&quot;&gt;&lt;span class=&quot;autocomment&quot;&gt;Pranala Menarik&lt;/span&gt;&lt;/span&gt;&lt;/p&gt;
&lt;table class=&quot;diff diff-contentalign-left diff-editfont-monospace&quot; data-mw=&quot;interface&quot;&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;tr class=&quot;diff-title&quot; lang=&quot;en&quot;&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;← Older revision&lt;/td&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;Revision as of 01:01, 14 August 2019&lt;/td&gt;
				&lt;/tr&gt;&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l991&quot; &gt;Line 991:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 991:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;==Pranala Menarik==&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;==Pranala Menarik==&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot;&gt; &lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins style=&quot;font-weight: bold; text-decoration: none;&quot;&gt;&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot;&gt; &lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins style=&quot;font-weight: bold; text-decoration: none;&quot;&gt;* [[Keras]]&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;/table&gt;</summary>
		<author><name>Onnowpurbo</name></author>
	</entry>
	<entry>
		<id>https://onnocenter.or.id/wiki/index.php?title=Keras:_Python_Keras_Text_Classification&amp;diff=56752&amp;oldid=prev</id>
		<title>Onnowpurbo: /* Conclusion */</title>
		<link rel="alternate" type="text/html" href="https://onnocenter.or.id/wiki/index.php?title=Keras:_Python_Keras_Text_Classification&amp;diff=56752&amp;oldid=prev"/>
		<updated>2019-08-14T00:56:28Z</updated>

		<summary type="html">&lt;p&gt;&lt;span dir=&quot;auto&quot;&gt;&lt;span class=&quot;autocomment&quot;&gt;Conclusion&lt;/span&gt;&lt;/span&gt;&lt;/p&gt;
&lt;table class=&quot;diff diff-contentalign-left diff-editfont-monospace&quot; data-mw=&quot;interface&quot;&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;tr class=&quot;diff-title&quot; lang=&quot;en&quot;&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;← Older revision&lt;/td&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;Revision as of 00:56, 14 August 2019&lt;/td&gt;
				&lt;/tr&gt;&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l970&quot; &gt;Line 970:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 970:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;==Conclusion==&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;==Conclusion==&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;There you have it: you have learned how to work with text classification with &lt;/del&gt;Keras, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;and we have gone from a &lt;/del&gt;bag-of-words &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;model with logistic regression to increasingly more advanced methods leading to &lt;/del&gt;convolutional neural networks.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Kita telah belajar cara bekerja dengan klasifikasi teks dengan &lt;/ins&gt;Keras, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dan kita telah beralih dari model &lt;/ins&gt;bag-of-words &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dengan regresi logistik menjadi metode yang semakin maju yang mengarah ke &lt;/ins&gt;convolutional neural networks.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;You should be now familiar with &lt;/del&gt;word embeddings, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;why they are useful&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;and also how to use &lt;/del&gt;pretrained word embeddings &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;for your &lt;/del&gt;training. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;You have also learned how to work with &lt;/del&gt;neural &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;networks and how to use &lt;/del&gt;hyperparameter &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;optimization to squeeze more performance out of your &lt;/del&gt;model.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Kita sekarang harus terbiasa dengan &lt;/ins&gt;word embeddings, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;mengapa mereka berguna&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dan juga bagaimana menggunakan &lt;/ins&gt;pretrained word embeddings &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;untuk &lt;/ins&gt;training &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;kita&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt; Kita juga telah belajar cara bekerja dengan &lt;/ins&gt;neural &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;network dan cara menggunakan optimasi &lt;/ins&gt;hyperparameter &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;untuk memeras lebih banyak kinerja dari &lt;/ins&gt;model &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;kita&lt;/ins&gt;.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;One big topic which we have not covered here left for another time was &lt;/del&gt;recurrent neural networks, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;more specifically &lt;/del&gt;LSTM &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;and &lt;/del&gt;GRU. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;Those are other powerful and popular tools to work with &lt;/del&gt;sequential &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;data like text or &lt;/del&gt;time series. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;Other interesting developments are currently in &lt;/del&gt;neural &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;networks that employ attention which are under active research and seem to be a promising next step since &lt;/del&gt;LSTM &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;tend to be heavy on the computation&lt;/del&gt;.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Satu topik besar yang belum kita bahas di sini dibiarkan lain waktu adalah &lt;/ins&gt;recurrent neural networks, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;lebih khusus &lt;/ins&gt;LSTM &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dan &lt;/ins&gt;GRU. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Itu adalah tool powerfull dan populer lainnya untuk bekerja dengan data &lt;/ins&gt;sequential &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;seperti teks atau &lt;/ins&gt;time series. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Perkembangan menarik lainnya saat ini dalam &lt;/ins&gt;neural &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;network yang menggunakan perhatian yang sedang dalam penelitian aktif dan tampaknya menjadi langkah berikutnya yang menjanjikan karena &lt;/ins&gt;LSTM &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;cenderung berat pada perhitungan&lt;/ins&gt;.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;You have now an understanding of a crucial cornerstone in &lt;/del&gt;natural language processing &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;which you can use for &lt;/del&gt;text classification &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;of all sorts&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;Sentiment analysis is the most prominent example for this&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;but this includes many other applications such as&lt;/del&gt;:&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Kita sekarang memiliki pemahaman tentang landasan penting dalam &lt;/ins&gt;natural language processing &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;yang dapat kita gunakan untuk semua jenis &lt;/ins&gt;text classification. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Analisis sentimen adalah contoh yang paling menonjol untuk ini&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;tetapi ini mencakup banyak aplikasi lain seperti&lt;/ins&gt;:&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;* Spam &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;detection in emails&lt;/del&gt;&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;* &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Deteksi &lt;/ins&gt;Spam &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;di email&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;* Automatic tagging &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;of texts&lt;/del&gt;&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;* Automatic tagging &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;pada text&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;* &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;Categorization of news articles with predefined topics&lt;/del&gt;&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;* &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Kategorisasi artikel berita untuk topik yang sudah di definisikan.&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;You can use this knowledge and the models that you have trained on an advanced project as in this &lt;/del&gt;tutorial &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;to employ sentiment analysis on a &lt;/del&gt;continuous stream &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;of &lt;/del&gt;twitter &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;data with &lt;/del&gt;Kibana &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;and &lt;/del&gt;Elasticsearch. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;You could also combine sentiment analysis or &lt;/del&gt;text classification &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;with &lt;/del&gt;speech recognition &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;like in this handy &lt;/del&gt;tutorial &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;using the &lt;/del&gt;SpeechRecognition &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;library in &lt;/del&gt;Python.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Kita dapat menggunakan pengetahuan ini dan model-model yang telah kita latih pada proyek lanjutan seperti dalam &lt;/ins&gt;tutorial &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;ini untuk menggunakan analisis sentimen pada  &lt;/ins&gt;continuous stream &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;data &lt;/ins&gt;twitter &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dengan &lt;/ins&gt;Kibana &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dan &lt;/ins&gt;Elasticsearch. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Kita juga bisa menggabungkan analisis sentimen atau  &lt;/ins&gt;text classification &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dengan &lt;/ins&gt;speech recognition &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;seperti dalam &lt;/ins&gt;tutorial &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;praktis ini menggunakan library &lt;/ins&gt;SpeechRecognition &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;di &lt;/ins&gt;Python.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;==Referensi==&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;==Referensi==&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;/table&gt;</summary>
		<author><name>Onnowpurbo</name></author>
	</entry>
	<entry>
		<id>https://onnocenter.or.id/wiki/index.php?title=Keras:_Python_Keras_Text_Classification&amp;diff=56751&amp;oldid=prev</id>
		<title>Onnowpurbo: /* Optimasi Hyperparameters */</title>
		<link rel="alternate" type="text/html" href="https://onnocenter.or.id/wiki/index.php?title=Keras:_Python_Keras_Text_Classification&amp;diff=56751&amp;oldid=prev"/>
		<updated>2019-08-14T00:48:37Z</updated>

		<summary type="html">&lt;p&gt;&lt;span dir=&quot;auto&quot;&gt;&lt;span class=&quot;autocomment&quot;&gt;Optimasi Hyperparameters&lt;/span&gt;&lt;/span&gt;&lt;/p&gt;
&lt;table class=&quot;diff diff-contentalign-left diff-editfont-monospace&quot; data-mw=&quot;interface&quot;&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;tr class=&quot;diff-title&quot; lang=&quot;en&quot;&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;← Older revision&lt;/td&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;Revision as of 00:48, 14 August 2019&lt;/td&gt;
				&lt;/tr&gt;&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l869&quot; &gt;Line 869:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 869:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;      return model&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;      return model&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;Next&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;you want to define the &lt;/del&gt;parameter grid &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;that you want to use in &lt;/del&gt;training. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;This consists of a &lt;/del&gt;dictionary &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;with each parameters named as in the previous function&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;The number of spaces on the &lt;/del&gt;grid &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;is &lt;/del&gt;3 * 3 * 1 * 1 * 1, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;where each of those numbers is the number of different choices for a given &lt;/del&gt;parameter.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Selanjutnya&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;kita ingin menentukan &lt;/ins&gt;parameter grid &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;yang ingin kita gunakan dalam &lt;/ins&gt;training. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Ini terdiri dari &lt;/ins&gt;dictionary &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dengan masing-masing parameter dalam fungsi sebelumnya&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Jumlah spasi di &lt;/ins&gt;grid &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;adalah &lt;/ins&gt;3 * 3 * 1 * 1 * 1, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;di mana masing-masing angka tersebut adalah jumlah pilihan yang berbeda untuk &lt;/ins&gt;parameter &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;yang diberikan&lt;/ins&gt;.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;You can see how this could get computationally expensive very quickly&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;but luckily both &lt;/del&gt;grid search &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;and &lt;/del&gt;random search &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;are embarrassingly parallel&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;and the classes come with an &lt;/del&gt;n_jobs &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;parameter that lets you test &lt;/del&gt;grid &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;spaces in parallel&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;The &lt;/del&gt;parameter &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;grid is initialized with the following &lt;/del&gt;dictionary:&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Kita dapat melihat bagaimana ini bisa menjadi komputasi yang mahal dengan sangat cepat&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;tetapi untungnya &lt;/ins&gt;grid search &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dan &lt;/ins&gt;random search &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;sebetulnya paralel&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dan class muncul dengan parameter &lt;/ins&gt;n_jobs &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;yang memungkinkan kita menguji ruang &lt;/ins&gt;grid &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;secara paralel&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Kotak &lt;/ins&gt;parameter &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;diinisialisasi dengan &lt;/ins&gt;dictionary &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;berikut&lt;/ins&gt;:&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  param_grid = dict(num_filters=[32, 64, 128],&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  param_grid = dict(num_filters=[32, 64, 128],&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l879&quot; &gt;Line 879:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 879:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;                    maxlen=[100])&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;                    maxlen=[100])&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;Now you are already ready to start running the &lt;/del&gt;random search. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;In this example we iterate over each &lt;/del&gt;data &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;set and then you want to preprocess the &lt;/del&gt;data &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;in the same way as previously&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;Afterwards you take the previous function and add it to the &lt;/del&gt;KerasClassifier wrapper class &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;including the number of epochs&lt;/del&gt;.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Sekarang kita sudah siap untuk mulai menjalankan &lt;/ins&gt;random search. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Dalam contoh ini kita mengulang setiap set &lt;/ins&gt;data &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dan kemudian kita ingin memproses ulang &lt;/ins&gt;data &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dengan cara yang sama seperti sebelumnya&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Setelah itu kita mengambil fungsi sebelumnya dan menambahkannya ke &lt;/ins&gt;KerasClassifier wrapper class &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;termasuk jumlah epoch.&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot;&gt; &lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Instance yang dihasilkan dan parameter grid kemudian digunakan sebagai estimator di RandomSearchCV class&lt;/ins&gt;.  &lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;The resulting instance and the parameter grid are then used as the estimator in the RandomSearchCV class. Additionally&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;you can choose the number of folds in the &lt;/del&gt;k-folds cross-validation, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;which is in this case &lt;/del&gt;4. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;You have seen most of the code in this snippet before in our previous examples&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;Besides the &lt;/del&gt;RandomSearchCV &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;and &lt;/del&gt;KerasClassifier, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;I have added a little block of code handling the evaluation&lt;/del&gt;:&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Selain itu&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;kita dapat memilih jumlah fold dalam &lt;/ins&gt;k-folds cross-validation, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;yang dalam kasus ini &lt;/ins&gt;4. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Kita telah melihat sebagian besar kode dalam cuplikan ini sebelumnya dalam contoh sebelumnya&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Selain &lt;/ins&gt;RandomSearchCV &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dan &lt;/ins&gt;KerasClassifier, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt; telah menambahkan sedikit kode untuk penanganan evaluasi&lt;/ins&gt;:&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  from keras.wrappers.scikit_learn import KerasClassifier&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  from keras.wrappers.scikit_learn import KerasClassifier&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l946&quot; &gt;Line 946:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 947:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;          f.write(output_string)&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;          f.write(output_string)&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;This takes a while which is a perfect chance to go outside to get some fresh air or even go on a hike&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;depending on how many models you want to run&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;Let’s take a look what we have got&lt;/del&gt;:&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Ini membutuhkan waktu yang merupakan kesempatan sempurna untuk pergi keluar untuk mendapatkan udara segar atau bahkan mendaki gunung&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;tergantung pada berapa banyak model yang ingin kita jalankan&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Mari kita lihat apa yang kita punya&lt;/ins&gt;:&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  Running amazon data set&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  Running amazon data set&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l963&quot; &gt;Line 963:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 964:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  Test Accuracy : 0.8384&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  Test Accuracy : 0.8384&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;Interesting&lt;/del&gt;! &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;For some reason the &lt;/del&gt;testing &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;accuracy is higher than the &lt;/del&gt;training &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;accuracy which might be because there is a large variance in the scores during &lt;/del&gt;cross-validation. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;We can see that we were still not able to break much through the dreaded &lt;/del&gt;80%, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;which seems to be a natural limit for this &lt;/del&gt;data &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;with its given size&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;Remember that we have a small &lt;/del&gt;data &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;set and &lt;/del&gt;convolutional neural networks &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;tend to perform the best with large &lt;/del&gt;data &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;sets&lt;/del&gt;.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Menarik&lt;/ins&gt;! &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Untuk beberapa hal maka akurasi &lt;/ins&gt;testing &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;lebih tinggi daripada akurasi &lt;/ins&gt;training &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;yang mungkin karena ada perbedaan besar dalam skor selama &lt;/ins&gt;cross-validation. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Kita dapat melihat bahwa kita masih tidak dapat menembus banyak melalui &lt;/ins&gt;80% &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;yang ditakuti&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;yang tampaknya menjadi batas alami untuk &lt;/ins&gt;data &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;ini dengan ukuran yang diberikan&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Ingat bahwa kita memiliki satu set &lt;/ins&gt;data &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;kecil dan &lt;/ins&gt;convolutional neural networks &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;cenderung melakukan yang terbaik dengan set &lt;/ins&gt;data &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;besar&lt;/ins&gt;.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;Another method for &lt;/del&gt;CV &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;is the &lt;/del&gt;nested cross-validation (&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;shown here&lt;/del&gt;) &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;which is used when the &lt;/del&gt;hyperparameters &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;also need to be optimized&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;This is used because the resulting &lt;/del&gt;non-nested CV &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;model has a &lt;/del&gt;bias &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;toward the &lt;/del&gt;data &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;set which can lead to an overly optimistic score&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;You see&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;when doing &lt;/del&gt;hyperparameter &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;optimization as we did in the previous example, we are picking the best hyperparameters for that specific &lt;/del&gt;training &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;set but this does not mean that these hyperparameters generalize the best&lt;/del&gt;.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Metode lain untuk &lt;/ins&gt;CV &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;adalah &lt;/ins&gt;nested cross-validation (&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;ditampilkan di sini&lt;/ins&gt;) &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;yang digunakan ketika &lt;/ins&gt;hyperparameters &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;juga perlu dioptimalkan&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Ini digunakan karena model &lt;/ins&gt;non-nested CV &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;yang dihasilkan memiliki &lt;/ins&gt;bias &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;terhadap kumpulan &lt;/ins&gt;data &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;yang dapat menyebabkan skor terlalu optimis&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Kita tahu&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;ketika melakukan optimasi hiperparameter seperti yang kita lakukan dalam contoh sebelumnya, kita memilih &lt;/ins&gt;hyperparameter &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;terbaik untuk set &lt;/ins&gt;training &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;tertentu tetapi ini tidak berarti bahwa hyperparameter ini menggeneralisasi yang terbaik&lt;/ins&gt;.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;==Conclusion==&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;==Conclusion==&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;/table&gt;</summary>
		<author><name>Onnowpurbo</name></author>
	</entry>
	<entry>
		<id>https://onnocenter.or.id/wiki/index.php?title=Keras:_Python_Keras_Text_Classification&amp;diff=56750&amp;oldid=prev</id>
		<title>Onnowpurbo: /* Hyperparameters Optimization */</title>
		<link rel="alternate" type="text/html" href="https://onnocenter.or.id/wiki/index.php?title=Keras:_Python_Keras_Text_Classification&amp;diff=56750&amp;oldid=prev"/>
		<updated>2019-08-14T00:38:30Z</updated>

		<summary type="html">&lt;p&gt;&lt;span dir=&quot;auto&quot;&gt;&lt;span class=&quot;autocomment&quot;&gt;Hyperparameters Optimization&lt;/span&gt;&lt;/span&gt;&lt;/p&gt;
&lt;table class=&quot;diff diff-contentalign-left diff-editfont-monospace&quot; data-mw=&quot;interface&quot;&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;tr class=&quot;diff-title&quot; lang=&quot;en&quot;&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;← Older revision&lt;/td&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;Revision as of 00:38, 14 August 2019&lt;/td&gt;
				&lt;/tr&gt;&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l841&quot; &gt;Line 841:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 841:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;CNN bekerja paling baik dengan training set yang besar di mana mereka dapat menemukan generalisasi di mana model sederhana seperti regresi logistik tidak akan mampu.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;CNN bekerja paling baik dengan training set yang besar di mana mereka dapat menemukan generalisasi di mana model sederhana seperti regresi logistik tidak akan mampu.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;==Hyperparameters &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;Optimization&lt;/del&gt;==&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;==&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Optimasi &lt;/ins&gt;Hyperparameters==&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;One crucial steps of &lt;/del&gt;deep &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;learning and working with &lt;/del&gt;neural &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;networks is &lt;/del&gt;hyperparameter &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;optimization&lt;/del&gt;.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Salah satu langkah penting dalam &lt;/ins&gt;deep &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;learnig dan bekerja dengan &lt;/ins&gt;neural &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;network adalah optimasi &lt;/ins&gt;hyperparameter.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;As you saw in the models that we have used so far&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;even with simpler ones&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;you had a large number of parameters to &lt;/del&gt;tweak &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;and choose from&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;Those parameters are called &lt;/del&gt;hyperparameters. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;This is the most time consuming part of &lt;/del&gt;machine learning &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;and sadly there are no one-fits-all solutions ready&lt;/del&gt;.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Seperti yang kita lihat dalam model yang telah kita gunakan sejauh ini&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;bahkan dengan model yang lebih sederhana&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;kita memiliki sejumlah besar parameter untuk di-&lt;/ins&gt;tweak &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dan dipilih&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Parameter tersebut disebut &lt;/ins&gt;hyperparameters. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Ini adalah bagian yang paling memakan waktu dari &lt;/ins&gt;machine learning &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dan sayangnya tidak ada solusi yang cocok untuk semua&lt;/ins&gt;.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;When you have a look at the competitions on &lt;/del&gt;Kaggle, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;one of the largest places to compete against other fellow &lt;/del&gt;data &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;scientists&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;you can see that many of the winning teams and models have gone through a lot of tweaking and experimenting until they reached their prime&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;So don’t get discouraged when it gets tough and you reach a plateau&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;but rather think about the ways you could optimize the &lt;/del&gt;model &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;or the &lt;/del&gt;data.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Ketika anda melihat kompetisi di &lt;/ins&gt;Kaggle, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;salah satu tempat terbesar untuk bersaing dengan sesama ilmuwan &lt;/ins&gt;data &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;lainnya&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;anda dapat melihat bahwa banyak tim dan model pemenang telah melalui banyak penyesuaian dan percobaan hingga mencapai puncaknya&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Jadi jangan berkecil hati ketika menjadi sulit dan anda mencapai dataran tinggi&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;tetapi pikirkan cara anda dapat mengoptimalkan &lt;/ins&gt;model &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;atau &lt;/ins&gt;data.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;One popular method for &lt;/del&gt;hyperparameter &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;optimization is &lt;/del&gt;grid search. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;What this method does is it takes lists of parameters and it runs the &lt;/del&gt;model &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;with each &lt;/del&gt;parameter &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;combination that it can find&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;It is the most thorough way but also the most computationally heavy way to do this&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;Another common way&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;random search&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;which you’ll see in action here&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;simply takes random combinations of parameters&lt;/del&gt;.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Salah satu metode populer untuk optimasi &lt;/ins&gt;hyperparameter &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;adalah &lt;/ins&gt;grid search. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Apa yang dilakukan metode ini adalah mengambil daftar parameter dan menjalankan &lt;/ins&gt;model &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dengan setiap kombinasi &lt;/ins&gt;parameter &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;yang dapat ditemukannya&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Ini adalah cara yang paling menyeluruh tetapi juga cara yang paling berat secara komputasi untuk melakukan ini&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Cara umum lain&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;pencarian acak&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;yang akan kita lihat dalam tindakan di sini&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;cukup mengambil kombinasi parameter secara acak&lt;/ins&gt;.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;In order to apply &lt;/del&gt;random search &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;with &lt;/del&gt;Keras, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;you will need to use the &lt;/del&gt;KerasClassifier &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;which serves as a wrapper for the &lt;/del&gt;scikit-learn API. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;With this wrapper you are able to use the various tools available with &lt;/del&gt;scikit-learn &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;like &lt;/del&gt;cross-validation. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;The class that you need is &lt;/del&gt;RandomizedSearchCV &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;which implements &lt;/del&gt;random search &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;with &lt;/del&gt;cross-validation. Cross-validation &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;is a way to validate the &lt;/del&gt;model &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;and take the whole &lt;/del&gt;data set &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;and separate it into multiple &lt;/del&gt;testing &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;and &lt;/del&gt;training &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;data sets&lt;/del&gt;.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Untuk menerapkan &lt;/ins&gt;random search &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dengan &lt;/ins&gt;Keras, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;kita harus menggunakan &lt;/ins&gt;KerasClassifier &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;yang berfungsi sebagai pembungkus untuk &lt;/ins&gt;scikit-learn API. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Dengan pembungkus ini kita dapat menggunakan berbagai tool yang tersedia dengan &lt;/ins&gt;scikit-learn &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;seperti &lt;/ins&gt;cross-validation. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Class yang kita butuhkan adalah &lt;/ins&gt;RandomizedSearchCV &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;yang mengimplementasikan &lt;/ins&gt;random search &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dengan &lt;/ins&gt;cross-validation. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt; &lt;/ins&gt;Cross-validation &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;adalah cara untuk memvalidasi &lt;/ins&gt;model &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dan mengambil seluruh rangkaian data dan memisahkannya menjadi beberapa &lt;/ins&gt;data set testing &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dan &lt;/ins&gt;training.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;There are various types of &lt;/del&gt;cross-validation. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;One type is the &lt;/del&gt;k-fold &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;cross-validation which you’ll see in this example&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;In this type the &lt;/del&gt;data set &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;is partitioned into k equal sized sets where one &lt;/del&gt;set &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;is used for &lt;/del&gt;testing &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;and the rest of the partitions are used for &lt;/del&gt;training. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;This enables you to run &lt;/del&gt;k &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;different runs&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;where each partition is once used as a &lt;/del&gt;testing &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;set&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;So&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;the higher &lt;/del&gt;k &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;is the more accurate the &lt;/del&gt;model &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;evaluation is&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;but the smaller each &lt;/del&gt;testing &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;set is&lt;/del&gt;.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Ada berbagai jenis &lt;/ins&gt;cross-validation. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Salah satu jenis adalah cross-validation &lt;/ins&gt;k-fold &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;yang akan kita lihat dalam contoh ini&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Dalam tipe ini set &lt;/ins&gt;data &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dipartisi ke dalam k &lt;/ins&gt;set &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;ukuran yang sama di mana satu &lt;/ins&gt;set &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;digunakan untuk &lt;/ins&gt;testing &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dan sisanya dari partisi digunakan untuk &lt;/ins&gt;training. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Ini memungkinkan kita untuk menjalankan &lt;/ins&gt;k &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dalam berbagai proses yang berbeda&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;di mana setiap partisi pernah digunakan sebagai set &lt;/ins&gt;testing. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Jadi&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;semakin tinggi &lt;/ins&gt;k &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;adalah semakin akurat &lt;/ins&gt;model &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;evaluasi&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;tetapi semakin kecil setiap set &lt;/ins&gt;testing&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;-nya&lt;/ins&gt;.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;First step for &lt;/del&gt;KerasClassifier &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;is to have a function that creates a &lt;/del&gt;Keras &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;model&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;We will use the previous &lt;/del&gt;model, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;but we will allow various parameters to be set for the &lt;/del&gt;hyperparameter &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;optimization&lt;/del&gt;:&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Langkah pertama untuk &lt;/ins&gt;KerasClassifier &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;adalah memiliki fungsi yang menciptakan model &lt;/ins&gt;Keras. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Kita akan menggunakan &lt;/ins&gt;model &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;sebelumnya&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;tetapi kita akan mengizinkan berbagai parameter ditetapkan untuk optimasi &lt;/ins&gt;hyperparameter:&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  def create_model(num_filters, kernel_size, vocab_size, embedding_dim, maxlen):&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  def create_model(num_filters, kernel_size, vocab_size, embedding_dim, maxlen):&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;/table&gt;</summary>
		<author><name>Onnowpurbo</name></author>
	</entry>
	<entry>
		<id>https://onnocenter.or.id/wiki/index.php?title=Keras:_Python_Keras_Text_Classification&amp;diff=56749&amp;oldid=prev</id>
		<title>Onnowpurbo: /* 1D Convolution (Image source) */</title>
		<link rel="alternate" type="text/html" href="https://onnocenter.or.id/wiki/index.php?title=Keras:_Python_Keras_Text_Classification&amp;diff=56749&amp;oldid=prev"/>
		<updated>2019-08-14T00:29:42Z</updated>

		<summary type="html">&lt;p&gt;&lt;span dir=&quot;auto&quot;&gt;&lt;span class=&quot;autocomment&quot;&gt;1D Convolution (Image source)&lt;/span&gt;&lt;/span&gt;&lt;/p&gt;
&lt;table class=&quot;diff diff-contentalign-left diff-editfont-monospace&quot; data-mw=&quot;interface&quot;&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;tr class=&quot;diff-title&quot; lang=&quot;en&quot;&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;← Older revision&lt;/td&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;Revision as of 00:29, 14 August 2019&lt;/td&gt;
				&lt;/tr&gt;&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l780&quot; &gt;Line 780:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 780:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;==1D Convolution (Image source)==&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;==1D Convolution (Image source)==&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;Now let’s have a look how you can use this &lt;/del&gt;network &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;in &lt;/del&gt;Keras. Keras &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;offers again various Convolutional layers which you can use for this task&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;The layer you’ll need is the &lt;/del&gt;Conv1D &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;layer&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;This layer has again various parameters to choose from&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;The ones you are interested in for now are the number of filters&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;the &lt;/del&gt;kernel &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;size&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;and the activation function&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;You can add this layer in between the &lt;/del&gt;Embedding &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;layer and the &lt;/del&gt;GlobalMaxPool1D &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;layer&lt;/del&gt;:&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Sekarang mari kita lihat bagaimana kita dapat menggunakan &lt;/ins&gt;network &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;ini di &lt;/ins&gt;Keras. Keras &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;menawarkan lagi berbagai lapisan konvolusional yang dapat kita gunakan untuk tugas ini&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Lapisan yang kita butuhkan adalah lapisan &lt;/ins&gt;Conv1D. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Lapisan ini memiliki lagi berbagai parameter untuk dipilih&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Yang kita minati saat ini adalah jumlah filter&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;ukuran &lt;/ins&gt;kernel, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dan fungsi aktivasi&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Kita bisa menambahkan lapisan ini di antara lapisan &lt;/ins&gt;Embedding &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dan lapisan &lt;/ins&gt;GlobalMaxPool1D:&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  embedding_dim = 100&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  embedding_dim = 100&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l795&quot; &gt;Line 795:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 795:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  model.summary()&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  model.summary()&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;The result will be as follows&lt;/del&gt;:&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Hasilnya&lt;/ins&gt;:&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  _________________________________________________________________&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  _________________________________________________________________&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l826&quot; &gt;Line 826:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 826:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  plot_history(history)&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  plot_history(history)&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;The result will be as follows&lt;/del&gt;:&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Hasilnya&lt;/ins&gt;:&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  Training Accuracy: 1.0000&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  Training Accuracy: 1.0000&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  Testing Accuracy:  0.7700&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  Testing Accuracy:  0.7700&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;loss &lt;/del&gt;accuracy convolution &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;model&lt;/del&gt;&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;[[File:Loss-&lt;/ins&gt;accuracy&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;-&lt;/ins&gt;convolution&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;-model4.png|center|400px|thumb]]&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;Accuracy and loss for convolutional neural network&lt;/del&gt;&lt;/div&gt;&lt;/td&gt;&lt;td colspan=&quot;2&quot;&gt; &lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;You can see that &lt;/del&gt;80% &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;accuracy seems to be tough hurdle to overcome with this &lt;/del&gt;data &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;set and a &lt;/del&gt;CNN &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;might not be well equipped&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;The reason for such a plateau might be that&lt;/del&gt;:&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Anda dapat melihat bahwa accuracy &lt;/ins&gt;80% &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;tampaknya merupakan rintangan yang sulit untuk diatasi dengan kumpulan &lt;/ins&gt;data &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;ini dan &lt;/ins&gt;CNN &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;mungkin tidak dilengkapi dengan baik&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Alasan untuk dataran tinggi accuracy seperti itu adalah&lt;/ins&gt;:&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;    There are not enough &lt;/del&gt;training &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;samples&lt;/del&gt;&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;* Tidak cukup banyak sample &lt;/ins&gt;training&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;    The data you have does not &lt;/del&gt;generalize &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;well&lt;/del&gt;&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;* Data yang kita miliki kurang &lt;/ins&gt;generalize&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;    Missing focus on tweaking the &lt;/del&gt;hyperparameters&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;* Gagal fokus saat mengatur &lt;/ins&gt;hyperparameters&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;CNNs work best with large &lt;/del&gt;training &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;sets where they are able to find generalizations where a simple &lt;/del&gt;model &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;like logistic regression won’t be able&lt;/del&gt;.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;CNN bekerja paling baik dengan &lt;/ins&gt;training &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;set yang besar di mana mereka dapat menemukan generalisasi di mana &lt;/ins&gt;model &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;sederhana seperti regresi logistik tidak akan mampu&lt;/ins&gt;.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;==Hyperparameters Optimization==&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;==Hyperparameters Optimization==&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;/table&gt;</summary>
		<author><name>Onnowpurbo</name></author>
	</entry>
	<entry>
		<id>https://onnocenter.or.id/wiki/index.php?title=Keras:_Python_Keras_Text_Classification&amp;diff=56747&amp;oldid=prev</id>
		<title>Onnowpurbo: /* Convolutional Neural Networks (CNN) */</title>
		<link rel="alternate" type="text/html" href="https://onnocenter.or.id/wiki/index.php?title=Keras:_Python_Keras_Text_Classification&amp;diff=56747&amp;oldid=prev"/>
		<updated>2019-08-14T00:24:33Z</updated>

		<summary type="html">&lt;p&gt;&lt;span dir=&quot;auto&quot;&gt;&lt;span class=&quot;autocomment&quot;&gt;Convolutional Neural Networks (CNN)&lt;/span&gt;&lt;/span&gt;&lt;/p&gt;
&lt;table class=&quot;diff diff-contentalign-left diff-editfont-monospace&quot; data-mw=&quot;interface&quot;&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;tr class=&quot;diff-title&quot; lang=&quot;en&quot;&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;← Older revision&lt;/td&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;Revision as of 00:24, 14 August 2019&lt;/td&gt;
				&lt;/tr&gt;&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l775&quot; &gt;Line 775:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 775:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;Pada gambar berikut, kita dapat melihat bagaimana konvolusi bekerja. Dimulai dengan menambahkan patch fitur input dengan ukuran kernel filter. Dengan patch ini, kita mengambil produk dot dari Weight filter yang dikalikan. Convnet satu dimensi tidak sama dengan terjemahan, yang berarti bahwa urutan tertentu dapat dikenali pada posisi yang berbeda. Ini dapat membantu untuk pola-pola tertentu dalam teks:&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;Pada gambar berikut, kita dapat melihat bagaimana konvolusi bekerja. Dimulai dengan menambahkan patch fitur input dengan ukuran kernel filter. Dengan patch ini, kita mengambil produk dot dari Weight filter yang dikalikan. Convnet satu dimensi tidak sama dengan terjemahan, yang berarti bahwa urutan tertentu dapat dikenali pada posisi yang berbeda. Ini dapat membantu untuk pola-pola tertentu dalam teks:&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot;&gt; &lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins style=&quot;font-weight: bold; text-decoration: none;&quot;&gt;&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot;&gt; &lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins style=&quot;font-weight: bold; text-decoration: none;&quot;&gt;[[File:1d-convolution.png|center|400px|thumb]]&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;==1D Convolution (Image source)==&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;==1D Convolution (Image source)==&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;/table&gt;</summary>
		<author><name>Onnowpurbo</name></author>
	</entry>
	<entry>
		<id>https://onnocenter.or.id/wiki/index.php?title=Keras:_Python_Keras_Text_Classification&amp;diff=56745&amp;oldid=prev</id>
		<title>Onnowpurbo: /* Convolutional Neural Networks (CNN) */</title>
		<link rel="alternate" type="text/html" href="https://onnocenter.or.id/wiki/index.php?title=Keras:_Python_Keras_Text_Classification&amp;diff=56745&amp;oldid=prev"/>
		<updated>2019-08-14T00:23:11Z</updated>

		<summary type="html">&lt;p&gt;&lt;span dir=&quot;auto&quot;&gt;&lt;span class=&quot;autocomment&quot;&gt;Convolutional Neural Networks (CNN)&lt;/span&gt;&lt;/span&gt;&lt;/p&gt;
&lt;table class=&quot;diff diff-contentalign-left diff-editfont-monospace&quot; data-mw=&quot;interface&quot;&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;tr class=&quot;diff-title&quot; lang=&quot;en&quot;&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;← Older revision&lt;/td&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;Revision as of 00:23, 14 August 2019&lt;/td&gt;
				&lt;/tr&gt;&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l770&quot; &gt;Line 770:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 770:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;CNN memiliki lapisan tersembunyi yang disebut lapisan konvolusional. Ketika kita membayangkan sebuah gambar, komputer harus berurusan dengan matrix dua dimensi angka dan oleh karena itu kita perlu beberapa cara untuk mendeteksi fitur dalam matrix ini. Lapisan konvolusional ini mampu mendeteksi tepi, sudut dan jenis tekstur lainnya yang menjadikannya tool yang istimewa. Lapisan convolutional terdiri dari beberapa filter yang digeser melintasi gambar dan dapat mendeteksi fitur-fitur tertentu.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;CNN memiliki lapisan tersembunyi yang disebut lapisan konvolusional. Ketika kita membayangkan sebuah gambar, komputer harus berurusan dengan matrix dua dimensi angka dan oleh karena itu kita perlu beberapa cara untuk mendeteksi fitur dalam matrix ini. Lapisan konvolusional ini mampu mendeteksi tepi, sudut dan jenis tekstur lainnya yang menjadikannya tool yang istimewa. Lapisan convolutional terdiri dari beberapa filter yang digeser melintasi gambar dan dapat mendeteksi fitur-fitur tertentu.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;Ini adalah inti dari tekniknya, proses konvolusi matematika. Dengan setiap lapisan konvolusional, network dapat mendeteksi pola yang lebih kompleks. Dalam Feature Visualization oleh Chris Olah kita bisa mendapatkan intuisi yang baik seperti apa fitur-fitur ini.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;Ini adalah inti dari tekniknya, proses konvolusi matematika. Dengan setiap lapisan konvolusional, network dapat mendeteksi pola yang lebih kompleks. Dalam Feature Visualization &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;https://distill.pub/2017/feature-visualization/ &lt;/ins&gt;oleh Chris Olah kita bisa mendapatkan intuisi yang baik seperti apa fitur-fitur ini.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;When you are working with sequential &lt;/del&gt;data, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;like text&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;you work with one dimensional convolutions&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;but the idea and the application stays the same&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;You still want to pick up on patterns in the sequence which become more complex with each added convolutional layer&lt;/del&gt;.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Saat kita bekerja dengan &lt;/ins&gt;data &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;sekuensial&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;seperti teks&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;kita bekerja dengan konvolusi satu dimensi&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;tetapi gagasan dan aplikasinya tetap sama&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Kita masih ingin menangkap pola dalam urutan yang menjadi lebih kompleks dengan setiap lapisan konvolusional yang ditambahkan&lt;/ins&gt;.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;In the next figure you can see how such a convolution works&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;It starts by taking a &lt;/del&gt;patch &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;of &lt;/del&gt;input &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;features with the size of the &lt;/del&gt;filter &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;kernel&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;With this &lt;/del&gt;patch &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;you take the &lt;/del&gt;dot &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;product of the multiplied weights of the &lt;/del&gt;filter. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;The one dimensional convnet is invariant to translations&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;which means that certain sequences can be recognized at a different position&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;This can be helpful for certain patterns in the text&lt;/del&gt;:&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Pada gambar berikut, kita dapat melihat bagaimana konvolusi bekerja&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Dimulai dengan menambahkan &lt;/ins&gt;patch &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;fitur &lt;/ins&gt;input &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dengan ukuran kernel &lt;/ins&gt;filter. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Dengan &lt;/ins&gt;patch &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;ini, kita mengambil produk &lt;/ins&gt;dot &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dari Weight &lt;/ins&gt;filter &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;yang dikalikan&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Convnet satu dimensi tidak sama dengan terjemahan&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;yang berarti bahwa urutan tertentu dapat dikenali pada posisi yang berbeda&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Ini dapat membantu untuk pola-pola tertentu dalam teks&lt;/ins&gt;:&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;one dimensional convolution&lt;/del&gt;&lt;/div&gt;&lt;/td&gt;&lt;td colspan=&quot;2&quot;&gt; &lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;==1D Convolution (Image source)==&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;==1D Convolution (Image source)==&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;/table&gt;</summary>
		<author><name>Onnowpurbo</name></author>
	</entry>
	<entry>
		<id>https://onnocenter.or.id/wiki/index.php?title=Keras:_Python_Keras_Text_Classification&amp;diff=56744&amp;oldid=prev</id>
		<title>Onnowpurbo: /* Convolutional Neural Networks (CNN) */</title>
		<link rel="alternate" type="text/html" href="https://onnocenter.or.id/wiki/index.php?title=Keras:_Python_Keras_Text_Classification&amp;diff=56744&amp;oldid=prev"/>
		<updated>2019-08-14T00:18:52Z</updated>

		<summary type="html">&lt;p&gt;&lt;span dir=&quot;auto&quot;&gt;&lt;span class=&quot;autocomment&quot;&gt;Convolutional Neural Networks (CNN)&lt;/span&gt;&lt;/span&gt;&lt;/p&gt;
&lt;table class=&quot;diff diff-contentalign-left diff-editfont-monospace&quot; data-mw=&quot;interface&quot;&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;tr class=&quot;diff-title&quot; lang=&quot;en&quot;&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;← Older revision&lt;/td&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;Revision as of 00:18, 14 August 2019&lt;/td&gt;
				&lt;/tr&gt;&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l770&quot; &gt;Line 770:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 770:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;CNN memiliki lapisan tersembunyi yang disebut lapisan konvolusional. Ketika kita membayangkan sebuah gambar, komputer harus berurusan dengan matrix dua dimensi angka dan oleh karena itu kita perlu beberapa cara untuk mendeteksi fitur dalam matrix ini. Lapisan konvolusional ini mampu mendeteksi tepi, sudut dan jenis tekstur lainnya yang menjadikannya tool yang istimewa. Lapisan convolutional terdiri dari beberapa filter yang digeser melintasi gambar dan dapat mendeteksi fitur-fitur tertentu.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;CNN memiliki lapisan tersembunyi yang disebut lapisan konvolusional. Ketika kita membayangkan sebuah gambar, komputer harus berurusan dengan matrix dua dimensi angka dan oleh karena itu kita perlu beberapa cara untuk mendeteksi fitur dalam matrix ini. Lapisan konvolusional ini mampu mendeteksi tepi, sudut dan jenis tekstur lainnya yang menjadikannya tool yang istimewa. Lapisan convolutional terdiri dari beberapa filter yang digeser melintasi gambar dan dapat mendeteksi fitur-fitur tertentu.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del class=&quot;diffchange diffchange-inline&quot;&gt;This is the very core of the technique&lt;/del&gt;, &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;the mathematical process of convolution&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;With each convolutional layer the &lt;/del&gt;network &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;is able to detect more complex patterns&lt;/del&gt;. &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;In the &lt;/del&gt;Feature Visualization &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;by &lt;/del&gt;Chris Olah &lt;del class=&quot;diffchange diffchange-inline&quot;&gt;you can get a good intuition what these features can look like&lt;/del&gt;.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Ini adalah inti dari tekniknya&lt;/ins&gt;, &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;proses konvolusi matematika&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Dengan setiap lapisan konvolusional, &lt;/ins&gt;network &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;dapat mendeteksi pola yang lebih kompleks&lt;/ins&gt;. &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;Dalam &lt;/ins&gt;Feature Visualization &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;oleh &lt;/ins&gt;Chris Olah &lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;kita bisa mendapatkan intuisi yang baik seperti apa fitur-fitur ini&lt;/ins&gt;.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;When you are working with sequential data, like text, you work with one dimensional convolutions, but the idea and the application stays the same. You still want to pick up on patterns in the sequence which become more complex with each added convolutional layer.&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;When you are working with sequential data, like text, you work with one dimensional convolutions, but the idea and the application stays the same. You still want to pick up on patterns in the sequence which become more complex with each added convolutional layer.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;/table&gt;</summary>
		<author><name>Onnowpurbo</name></author>
	</entry>
</feed>