apaam/mlpack
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

updated niche/functional

Browse Source
This commit is contained in:
tekhnofiend
2008-02-19 16:55:40 +00:00
parent c8a16a96f6
commit 6e2a04cb3d
2 changed files with 22 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files
fastlib/u/niche/functional/p300_funcica.m
-1
View File
@@ -96,7 +96,6 @@ centered_data_coef = ...
repmat(getcoef(mean_result.meanfd), 1, size(getcoef(myfd), 2));
centered_myfd = fd(centered_data_coef, mybasis);
basis_curves = eval_basis(t, mybasis);
centered_data_curves = basis_curves * getcoef(centered_myfd);
fastlib/u/niche/functional/test_genes.m
+22 -11
View File
@@ -31,26 +31,37 @@ mybasis = create_bspline_basis([min(t) max(t)], p, 4);
basis_curves = eval_basis(0:1:119, mybasis);
basis_inner_products = full(eval_penalty(mybasis, int2Lfd(0)));
myfd_data = data2fd(data, t, mybasis);
myfd = data2fd(data, t, mybasis);
mean_result = pca_fd(myfd, 0);
centered_data_coef = ...
getcoef(myfd) - ...
repmat(getcoef(mean_result.meanfd), 1, size(getcoef(myfd), 2));
centered_myfd = fd(centered_data_coef, mybasis);
centered_data_curves = basis_curves * getcoef(centered_myfd);
lambda = 0;
myfdPar = fdPar(mybasis, 2, lambda);
pca_results = pca_fd(centered_myfd, p, myfdPar);
fprintf('calling funcica\n');
[ic_curves, ic_coef, Y, pc_coef, pc_curves, pc_scores, W] = ...
funcica(t, myfd_data, p, basis_curves, myfdPar, basis_inner_products);
[ic_curves, ic_coef, ic_scores, ...
pc_coef, pc_curves, pc_scores, W] = ...
funcica(t, centered_myfd, p, basis_curves, myfdPar, ...
basis_inner_products);
fprintf('funcica returned\n');
p_small = size(ic_coef, 2);
pc_curves = basis_curves * pc_coef;
ic_curves = pc_curves * W';
pc_coef = pc_coef';
for i = 1:size(ic_curves,2)
scale_up_factor = ...
1 / sqrt(l2_fnorm(t, ic_curves(:,i), ic_curves(:,i)));
ic_curves(:,i) = scale_up_factor * ic_curves(:,i);
ic_scores(i,:) = scale_up_factor * ic_scores(i,:);
end
%{
% rescale the utilized parts of pc_coef such that
% the pc_curves square integrate to 1
@@ -79,7 +90,6 @@ h_ic_sum = sum(h_ic);
% discriminant analysis using pc features %
% pc_scores is d x N
@@ -100,3 +110,4 @@ for i=1:size(svm_data, 1)
svm_labels([1:(i-1) (i+1):end]));
ypred(i) = svmlfwd(latestSVM, svm_data(i,:), svm_labels(i));
end
%}