I want a code for kernel PCA

hello, I am looking for a site where I can find a program reduction dimentionality kernel PCA think you

Answer 1

[Shashank Prasanna]

MATLAB currently doesn't offer this as a part of the statistics toolbox. In the past I've used this and it worked really well for me:

http://homepage.tudelft.nl/19j49/Matlab_Toolbox_for_Dimensionality_Reduction.html

[BANI tita]

Hello, thank you for your reply, I tried to execute the function kernel pca but she give a following error ,pray you help me find a solution.

%%=============================

??? Subscripted assignment dimension mismatch.

Error in ==> eigs at 282 workd(:,cols(2)) = applyOP(workd(:,cols(1)));

Error in ==> kernel_pca at 77 [V, L] = eigs(@(v)kernel_function(v, X', 1, kernel, param1, param2, 'Normal'), size(X, 1), no_dims, 'LM', options);

Error in ==> compute_mapping at 348 else [mappedA, mapping] = kernel_pca(A, no_dims, varargin{:}); end

Error in ==> test_kernel_pca at 12 mappedX = compute_mapping(X, 'KernelPCA', no_dims, 7);

%======================
  % load baseAT.mat;
X=PA;
no_dims =10;
labels=lblsa;
mappedX = compute_mapping(X, 'KernelPCA', no_dims, 7);
%%============

[Walter Roberson]

what is size(X) ?

[BANI tita]

size(x)=(5185,220)

[Shashank Prasanna]

try clear all

try some of the trouble shooting steps mentioned in the website:

http://homepage.tudelft.nl/19j49/Matlab_Toolbox_for_Dimensionality_Reduction.html

make sure you have the relevant files on the path

[BANI tita]

hello,

here is the program I run but still it gives me same error

 %%%-----------------------------------

load baseAT.mat;
    X=PA;
    labels=lblsa;
  figure, scatter3(X(:,1), X(:,2), X(:,3), 5, labels); title('Original dataset'), drawnow
  no_dims = round(intrinsic_dim(X, 'EigValue'));
  disp(['MLE estimate of intrinsic dimensionality: ' num2str(no_dims)]);
  mappedX = compute_mapping(X, 'KernelPCA', no_dims, 7);  
  figure, scatter(mappedX(:,1), mappedX(:,2), 5, labels); title('Result of dimensionality reduction'), drawnow

[Hariprasad Kannan]

The kernel pca code doesn't address the pre-image problem. It gives location of the points in the feature space.