pca

chapter06/demo.m

@@ -46,6 +46,9 @@
 % maxabsdiff(y_kn,y_lin)
 % maxabsdiff(sigma_kn,sigma_lin)
 % maxabsdiff(p_kn,p_lin)
+%% kernel PCA with linear kernel is PCA
+% clear; close all;
+% n = 100;
 %% test case for knCenter
 % kn = @knGauss;
 % X=rand(2,100);

chapter06/knKmeans.m

@@ -1,7 +1,7 @@
 function [label, energy, model] = knKmeans(X, init, kn)
 % Perform kernel k-means clustering.
 %   K: nxn kernel matrix
-%   k: number of cluster
+%   init: either number of clusters or initial label
 % Reference: Kernel Methods for Pattern Analysis
 % by John Shawe-Taylor, Nello Cristianini
 % Written by Mo Chen ([email protected]).

chapter06/knPca.m

@@ -1,5 +1,9 @@
-function [ R, Z, err ] = knPca( X, d, kn )
+function [ R, Z, err, model] = knPca(X, d, kn)
 % Kernel PCA
+%   X: dxn data matrix 
+%   d: target dimension
+%   kn: kernel function
+% Written by Mo Chen ([email protected]).
 if nargin < 3
     kn = @knGauss;
 end

chapter06/knPcaPred.m

@@ -1,7 +1,9 @@
-function [ output_args ] = knPcaPred( input_args )
-%KNPCAPRED Summary of this function goes here
-%   Detailed explanation goes here
+function X = knPcaPred(model, Xt)
+% Prediction for kernel PCA
+%   model: trained model structure
+%   X: d x n testing data
+%   t (optional): 1 x n testing response
+% Written by Mo Chen ([email protected]).
 
 
-end