摘要:
% LOAD DATASET AND DO PATH FOLLOWING WITH CROSS-VALIDATIONfunction PA=MKLEARNING2(x,y,dd)% x=feature_all;% y=dmos_all;% dd=2;% simulation parametersntotal = 1000; % total number of points - useful for subsampling datasetproptrain = .8; % proportion of training datansplits = dd; % number of splitssee 阅读全文
摘要:
function path = follow_entire_path(Ks,y,loss,ds,path_params,Ks_test,ytest)% FOLLOW THE ENTIRE PATH OF REGULARIZATION%% INPUT% Ks : kernel matrices% y : response% loss : loss.type= 'regression' or 'logistic'% ds : weights of the block 1-norm% path_params : parameters of the path (defi 阅读全文