digui.rar_recursive features_svm特征选择_特征选择 SVM_特征选择matlab_递归特征消除

% total number of samples 345 % input 230, output 115 % normalization: input [-1,1], output [0,1] function [P, T, TV] = liver_data %Obtain Random P, T load liver.dat; dataset = liver; rand_sequence=randperm(size(dataset,1)); temp_dataset=dataset; dataset=temp_dataset(rand_sequence, :); % for i=1:size(dataset,2)-1 % if max(dataset(:,i))~=min(dataset(:,i)) % dataset(:,i)=((dataset(:,i)-min(dataset(:,i)))/(max(dataset(:,i))-min(dataset(:,i))))*2-1; % else % dataset(:,i)=0; % end % end % T=dataset(:,size(dataset,2)); % if max(T)~=min(T) % T=(T-min(T))/(max(T)-min(T)); % else % T=ones(size(T))*0.5; % end % dataset(:,size(dataset,2))=T; P1=dataset(1:230,1:6); T1=dataset(1:230,7:7); % P=P1'; % T=T1'; %Obtain Random Test Matrix X=dataset(231:size(dataset,1),1:6); Y=dataset(231:size(dataset,1),7:7); % TV.P=X'; % TV.T=Y'; fid = fopen('liver_train','w'); for i=1:size(P1,1) fprintf(fid,'%2.8f ',T1(i,1)); for j=1:size(P1,2) % fprintf(fid,' %d:%2.8f',j, P1(i,j)); % for SVM fprintf(fid,' %2.8f', P1(i,j)); % for ELM end fprintf(fid,'\n'); end fclose(fid); fid = fopen('liver_test','w'); for i=1:size(X,1) fprintf(fid,'%2.8f ',Y(i,1)); for j=1:size(X,2) % fprintf(fid,' %d:%2.8f',j, X(i,j)); % for SVM fprintf(fid,' %2.8f', X(i,j)); % for ELM end fprintf(fid,'\n'); end fclose(fid);