Code Fragments Provided in the Book

5.1 Matlab code normalising a kernel matrix. page 113
5.2 Matlab code for centering a kernel matrix. 116
5.3 Matlab code for simple novelty detection algorithm. 118
5.4 Matlab code for performing incomplete Cholesky decomposition or dual partial Gram�Schmidt orthogonalisation. 129
5.5 Matlab code for standardising data. 131
5.6 Kernel Fisher discriminant algorithm 137
6.1 Matlab code for kernel PCA algorithm. 152
6.2 Pseudocode for the whitening algorithm. 156
6.3 Pseudocode for the kernel CCA algorithm. 175
6.4 Pseudocode for dual principal components regression. 179
6.5 Pseudocode for PLS feature extraction. 182
6.6 Pseudocode for the primal PLS algorithm. 186
6.7 Matlab code for the primal PLS algorithm. 187
6.8 Pseudocode for the kernel PLS algorithm. 191
6.9 Matlab code for the dual PLS algorithm. 192
7.1 Pseudocode for computing the minimal hypersphere. 199
7.2 Pseudocode for soft hypersphere minimisation. 205
7.3 Pseudocode for the soft hypersphere. 208
7.4 Pseudocode for the hard margin SVM. 215
7.5 Pseudocode for the alternative version of the hard SVM. 218
7.6 Pseudocode for 1-norm soft margin SVM. 223
7.7 Pseudocode for the soft margin SVM. 225
7.8 Pseudocode for the 2-norm SVM. 229
7.9 Pseudocode for 2-norm support vector regression. 237
7.10 Pseudocode for 1-norm support vector regression. 238
7.11 Pseudocode for new SVR. 240
7.12 Pseudocode for the kernel perceptron algorithm. 242
7.13 Pseudocode for the kernel adatron algorithm. 247
7.14 Pseudocode for the on-line support vector regression. 249
8.1 Pseudocode for the soft ranking algorithm. 259
8.2 Pseudocode for on-line ranking. 262
8.3 Matlab code to perform k-means clustering. 275
8.4 Matlab code to implementing low-dimensional visualisation. 285
9.1 Pseudocode for ANOVA kernel. 301
9.2 Pseudocode for simple graph kernels. 308
11.1 Pseudocode for the all-non-contiguous subsequences kernel. 356
11.2 Pseudocode for the fixed length subsequences kernel. 359
11.3 Pseudocode for the gap-weighted subsequences kernel. 369
11.4 Pseudocode for trie-based implementation of spectrum kernel. 374
11.5 Pseudocode for the trie-based implementation of the mismatch kernel. 378
11.6 Pseudocode for trie-based restricted gap-weighted subsequences kernel. 381
11.7 Pseudocode for the co-rooted subtree kernel. 387
11.8 Pseudocode for the all-subtree kernel. 389
12.1 Pseudocode for the fixed length HMM kernel. 409
12.2 Pseudocode for the pair HMM kernel. 415
12.3 Pseudocode for the hidden tree model kernel. 420
12.4 Pseudocode to compute the Fisher scores for the fixed length Markov model Fisher kernel. 435