Difference between revisions of "Orange: Rank"

From OnnoWiki
Jump to navigation Jump to search
 
(9 intermediate revisions by the same user not shown)
Line 2: Line 2:
  
  
−
Ranking of attributes in classification or regression datasets.
+
Widget Rank menghitung pemeringkatan atribut dalam klasifikasi atau regresi data.
  
−
Inputs
+
==Input==
  
−
    Data: input dataset
+
Data: input dataset
 +
Scorer: models for feature scoring
  
−
    Scorer: models for feature scoring
+
==Output==
  
−
Outputs
+
Reduced Data: dataset with selected attributes
  
−
    Reduced Data: dataset with selected attributes
+
Widget Rank mempertimbangkan kumpulan data berlabel class (klasifikasi atau regresi) dan menilai atribut sesuai dengan korelasinya dengan class. Peringkat juga menerima model untuk penilaian, seperti linear regression, logistic regression, random forest, SGD, dll.
  
−
The Rank widget considers class-labeled datasets (classification or regression) and scores the attributes according to their correlation with the class. Rank accepts also models for scoring, such as linear regression, logistic regression, random forest, SGD, etc.
+
[[File:Rank-stamped.png|center|600px|thumb]]
  
−
[[File:Rank-stamped.png|center|200px|thumb]]
+
* Select attributes from the data table.
 +
* Data table with attributes (rows) and their scores by different scoring methods (columns)
 +
* Produce a report.
 +
* If ‘Send Automatically’ is ticked, the widget automatically communicates changes to other widgets.
  
−
    Select attributes from the data table.
+
==Scoring method==
  
−
    Data table with attributes (rows) and their scores by different scoring methods (columns)
+
Information Gain: the expected amount of information (reduction of entropy)
  
−
    Produce a report.
+
* Gain Ratio: a ratio of the information gain and the attribute’s intrinsic information, which reduces the bias towards multivalued features that occurs in information gain
 +
* Gini: the inequality among values of a frequency distribution
 +
* ANOVA: the difference between average vaules of the feature in different classes
 +
* Chi2: dependence between the feature and the class as measure by the chi-square statistic
 +
* ReliefF: the ability of an attribute to distinguish between classes on similar data instances
 +
* FCBF (Fast Correlation Based Filter): entropy-based measure, which also identifies redundancy due to pairwise correlations between features
  
−
    If ‘Send Automatically’ is ticked, the widget automatically communicates changes to other widgets.
+
Selain itu, kita dapat menghubungkan learner tertentu yang memungkinkan penilaian feature berdasarkan seberapa penting mereka dalam model yang dibuat learner (mis. Linear Regression / Logistic Regression, Random Forest, SGD).
  
−
Scoring methods
+
==Example: Attribute Ranking and Selection==
  
−
    Information Gain: the expected amount of information (reduction of entropy)
+
Di bawah ini, kita menggunakan widget Rank segera setelah widget File untuk mengurangi set atribut data dan hanya menyertakan yang paling informatif:
  
−
    Gain Ratio: a ratio of the information gain and the attribute’s intrinsic information, which reduces the bias towards multivalued features that occurs in information gain
+
[[File:Rank-Select-Schema.png|center|600px|thumb]]
−
    Gini: the inequality among values of a frequency distribution
 
−
    ANOVA: the difference between average vaules of the feature in different classes
 
−
    Chi2: dependence between the feature and the class as measure by the chi-square statistic
 
−
    ReliefF: the ability of an attribute to distinguish between classes on similar data instances
 
−
    FCBF (Fast Correlation Based Filter): entropy-based measure, which also identifies redundancy due to pairwise correlations between features
 
  
−
Additionally, you can connect certain learners that enable scoring the features according to how important they are in models that the learners build (e.g. Linear Regression / Logistic Regression, Random Forest, SGD).
+
Perhatikan bagaimana widget Rank menghasilkan set data yang hanya menyertakan atribut dengan skor terbaik:
−
Example: Attribute Ranking and Selection
 
  
−
Below, we have used the Rank widget immediately after the File widget to reduce the set of data attributes and include only the most informative ones:
+
[[File:Rank-Select-Widgets.png|center|600px|thumb]]
−
 
 
−
[[File:Rank-Select-Schema.png|center|200px|thumb]]
 
−
 
 
−
 
 
−
Notice how the widget outputs a dataset that includes only the best-scored attributes:
 
−
 
 
−
[[File:Rank-Select-Widgets.png|center|200px|thumb]]
 
  
 
==Example: Feature Subset Selection for Machine Learning==
 
==Example: Feature Subset Selection for Machine Learning==
  
−
What follows is a bit more complicated example. In the workflow below, we first split the data into a training set and a test set. In the upper branch, the training data passes through the Rank widget to select the most informative attributes, while in the lower branch there is no feature selection. Both feature selected and original datasets are passed to their own Test & Score widgets, which develop a Naive Bayes classifier and score it on a test set.
+
Contoh yang lebih rumit dari widget Rank, dalam workflow di bawah ini, kita pertama-tama membagi data menjadi satu set train dan satu set test. Di cabang atas, data training melewatikan ke widget Rank untuk memilih atribut yang paling informatif, sedangkan di cabang bawah tidak ada pemilihan feature. Baik set data feature yang dipilih dan original dilewatkan ke widget Test & Score mereka sendiri, yang mengembangkan classifier widget Naive Bayes dan menilainya pada set test.
−
 
 
−
[[File:Rank-and-Test.png|center|200px|thumb]]
 
−
 
 
−
For datasets with many features, a naive Bayesian classifier feature selection, as shown above, would often yield a better predictive accuracy.
 
−
 
 
  
 +
Untuk dataset dengan banyak fitur, pemilihan feature classifier Naive Bayes, seperti yang ditunjukkan di gambar berikut, seringkali menghasilkan akurasi prediksi yang lebih baik.
  
 +
[[File:Rank-and-Test.png|center|600px|thumb]]
  
 
==Referensi==
 
==Referensi==

Latest revision as of 09:56, 20 April 2020

Sumber: https://docs.biolab.si//3/visual-programming/widgets/data/rank.html


Widget Rank menghitung pemeringkatan atribut dalam klasifikasi atau regresi data.

Input

Data: input dataset
Scorer: models for feature scoring

Output

Reduced Data: dataset with selected attributes

Widget Rank mempertimbangkan kumpulan data berlabel class (klasifikasi atau regresi) dan menilai atribut sesuai dengan korelasinya dengan class. Peringkat juga menerima model untuk penilaian, seperti linear regression, logistic regression, random forest, SGD, dll.

Rank-stamped.png
  • Select attributes from the data table.
  • Data table with attributes (rows) and their scores by different scoring methods (columns)
  • Produce a report.
  • If ‘Send Automatically’ is ticked, the widget automatically communicates changes to other widgets.

Scoring method

Information Gain: the expected amount of information (reduction of entropy)

  • Gain Ratio: a ratio of the information gain and the attribute’s intrinsic information, which reduces the bias towards multivalued features that occurs in information gain
  • Gini: the inequality among values of a frequency distribution
  • ANOVA: the difference between average vaules of the feature in different classes
  • Chi2: dependence between the feature and the class as measure by the chi-square statistic
  • ReliefF: the ability of an attribute to distinguish between classes on similar data instances
  • FCBF (Fast Correlation Based Filter): entropy-based measure, which also identifies redundancy due to pairwise correlations between features

Selain itu, kita dapat menghubungkan learner tertentu yang memungkinkan penilaian feature berdasarkan seberapa penting mereka dalam model yang dibuat learner (mis. Linear Regression / Logistic Regression, Random Forest, SGD).

Example: Attribute Ranking and Selection

Di bawah ini, kita menggunakan widget Rank segera setelah widget File untuk mengurangi set atribut data dan hanya menyertakan yang paling informatif:

Rank-Select-Schema.png

Perhatikan bagaimana widget Rank menghasilkan set data yang hanya menyertakan atribut dengan skor terbaik:

Rank-Select-Widgets.png

Example: Feature Subset Selection for Machine Learning

Contoh yang lebih rumit dari widget Rank, dalam workflow di bawah ini, kita pertama-tama membagi data menjadi satu set train dan satu set test. Di cabang atas, data training melewatikan ke widget Rank untuk memilih atribut yang paling informatif, sedangkan di cabang bawah tidak ada pemilihan feature. Baik set data feature yang dipilih dan original dilewatkan ke widget Test & Score mereka sendiri, yang mengembangkan classifier widget Naive Bayes dan menilainya pada set test.

Untuk dataset dengan banyak fitur, pemilihan feature classifier Naive Bayes, seperti yang ditunjukkan di gambar berikut, seringkali menghasilkan akurasi prediksi yang lebih baik.

Rank-and-Test.png

Referensi

Pranala Menarik

Retrieved from "https://onnocenter.or.id/wiki/index.php?title=Orange:_Rank&oldid=61315"