Difference between revisions of "Orange: Manifold Learning"

From OnnoWiki
Jump to navigation Jump to search
 
(5 intermediate revisions by the same user not shown)
Line 1: Line 1:
 
Sumber: https://docs.biolab.si//3/visual-programming/widgets/unsupervised/manifoldlearning.html
 
Sumber: https://docs.biolab.si//3/visual-programming/widgets/unsupervised/manifoldlearning.html
  
 +
Widget Manifold Learning dapat melakukan pengurangan / reduksi dimensi secara nonlinear.
  
 +
==Input==
  
−
Nonlinear dimensionality reduction.
+
Data: input dataset
  
−
Inputs
+
==Output==
  
−
    Data: input dataset
+
Transformed Data: dataset with reduced coordinates
  
−
Outputs
+
Widget Manifold Learning adalah teknik yang menemukan manifold non-linear dalam ruang dimensi yang lebih tinggi. Widget Manifold Learning kemudian menampilkan koordinat baru yang sesuai dengan ruang dua dimensi. Data tersebut dapat kemudian divisualisasikan dengan widget Scatter Plot atau widget visualisasi lainnya.
−
 
 
−
    Transformed Data: dataset with reduced coordinates
 
−
 
 
−
Manifold Learning is a technique which finds a non-linear manifold within the higher-dimensional space. The widget then outputs new coordinates which correspond to a two-dimensional space. Such data can be later visualized with Scatter Plot or other visualization widgets.
 
  
 
[[File:Manifold-learning-stamped.png|center|200px|thumb]]
 
[[File:Manifold-learning-stamped.png|center|200px|thumb]]
  
  
−
    Method for manifold learning:
+
* Method for manifold learning:
−
        t-SNE
+
** t-SNE
−
        MDS, see also MDS widget
+
** MDS, see also MDS widget
−
        Isomap
+
** Isomap
−
        Locally Linear Embedding
+
** Locally Linear Embedding
−
        Spectral Embedding
+
** Spectral Embedding
−
    Set parameters for the method:
+
* Set parameters for the method:
−
        t-SNE (distance measures):
+
** t-SNE (distance measures):
−
            Euclidean distance
+
*** Euclidean distance
−
            Manhattan
+
*** Manhattan
−
            Chebyshev
+
*** Chebyshev
−
            Jaccard
+
*** Jaccard
−
            Mahalanobis
+
*** Mahalanobis
−
            Cosine
+
*** Cosine
−
        MDS (iterations and initialization):
+
** MDS (iterations and initialization):
−
            max iterations: maximum number of optimization interactions
+
*** max iterations: maximum number of optimization interactions
−
            initialization: method for initialization of the algorithm (PCA or random)
+
*** initialization: method for initialization of the algorithm (PCA or random)
−
        Isomap:
+
** Isomap:
−
            number of neighbors
+
*** number of neighbors
−
        Locally Linear Embedding:
+
** Locally Linear Embedding:
−
            method:
+
*** method:
−
                standard
+
**** standard
−
                modified
+
**** modified
−
                hessian eigenmap
+
**** hessian eigenmap
−
                local
+
**** local
−
            number of neighbors
+
*** number of neighbors
−
            max iterations
+
*** max iterations
−
        Spectral Embedding:
+
** Spectral Embedding:
−
            affinity:
+
*** affinity:
−
                nearest neighbors
+
**** nearest neighbors
−
                RFB kernel
+
**** RFB kernel
−
    Output: the number of reduced features (components).
+
* Output: the number of reduced features (components).
−
    If Apply automatically is ticked, changes will be propagated automatically. Alternatively, click Apply.
+
* If Apply automatically is ticked, changes will be propagated automatically. Alternatively, click Apply.
−
    Produce a report.
+
* Produce a report.
  
−
Manifold Learning widget produces different embeddings for high-dimensional data.
+
Widget Manifold Learning menghasilkan embeddings yang berbeda untuk data high-dimensional.
  
−
[[File:Collage-manifold.png|center|200px|thumb]]
+
[[File:Collage-manifold.png|center|600px|thumb]]
  
−
From left to right, top to bottom: t-SNE, MDS, Isomap, Locally Linear Embedding and Spectral Embedding.
+
Dari kiri ke kanan, atas ke bawah: t-SNE, MDS, Isomap, Locally Linear Embedding dan Spectral Embedding.
  
 
==Contoh==
 
==Contoh==
  
−
Manifold Learning widget transforms high-dimensional data into a lower dimensional approximation. This makes it great for visualizing datasets with many features. We used voting.tab to map 16-dimensional data onto a 2D graph. Then we used Scatter Plot to plot the embeddings.
+
Widget Manifold Learning mengubah data dimensi tinggi menjadi pendekatan dimensi yang lebih rendah. Ini membuatnya bagus untuk memvisualisasikan dataset dengan banyak fitur. Kita menggunakan voting.tab dari widget File untuk memetakan data 16 dimensi ke grafik 2D menggunakan widget Manifold Learning. Lalu kita menggunakan widget Scatter Plot untuk memplot embeddings.
−
 
 
−
[[File:Manifold-learning-example.png|center|200px|thumb]]
 
  
 +
[[File:Manifold-learning-example.png|center|600px|thumb]]
  
 
==Referensi==
 
==Referensi==

Latest revision as of 10:55, 15 April 2020

Sumber: https://docs.biolab.si//3/visual-programming/widgets/unsupervised/manifoldlearning.html

Widget Manifold Learning dapat melakukan pengurangan / reduksi dimensi secara nonlinear.

Input

Data: input dataset

Output

Transformed Data: dataset with reduced coordinates

Widget Manifold Learning adalah teknik yang menemukan manifold non-linear dalam ruang dimensi yang lebih tinggi. Widget Manifold Learning kemudian menampilkan koordinat baru yang sesuai dengan ruang dua dimensi. Data tersebut dapat kemudian divisualisasikan dengan widget Scatter Plot atau widget visualisasi lainnya.

Manifold-learning-stamped.png


  • Method for manifold learning:
    • t-SNE
    • MDS, see also MDS widget
    • Isomap
    • Locally Linear Embedding
    • Spectral Embedding
  • Set parameters for the method:
    • t-SNE (distance measures):
      • Euclidean distance
      • Manhattan
      • Chebyshev
      • Jaccard
      • Mahalanobis
      • Cosine
    • MDS (iterations and initialization):
      • max iterations: maximum number of optimization interactions
      • initialization: method for initialization of the algorithm (PCA or random)
    • Isomap:
      • number of neighbors
    • Locally Linear Embedding:
      • method:
        • standard
        • modified
        • hessian eigenmap
        • local
      • number of neighbors
      • max iterations
    • Spectral Embedding:
      • affinity:
        • nearest neighbors
        • RFB kernel
  • Output: the number of reduced features (components).
  • If Apply automatically is ticked, changes will be propagated automatically. Alternatively, click Apply.
  • Produce a report.

Widget Manifold Learning menghasilkan embeddings yang berbeda untuk data high-dimensional.

Collage-manifold.png

Dari kiri ke kanan, atas ke bawah: t-SNE, MDS, Isomap, Locally Linear Embedding dan Spectral Embedding.

Contoh

Widget Manifold Learning mengubah data dimensi tinggi menjadi pendekatan dimensi yang lebih rendah. Ini membuatnya bagus untuk memvisualisasikan dataset dengan banyak fitur. Kita menggunakan voting.tab dari widget File untuk memetakan data 16 dimensi ke grafik 2D menggunakan widget Manifold Learning. Lalu kita menggunakan widget Scatter Plot untuk memplot embeddings.

Manifold-learning-example.png

Referensi

Pranala Menarik

Retrieved from "https://onnocenter.or.id/wiki/index.php?title=Orange:_Manifold_Learning&oldid=61231"