KFDA 核fisher

Feature scaling for kernel Fisher discriminant analysis using leave-one-out cross validation. FS-KFDA is a package for implementing feature scaling for kernel fisher discriminant analysis.-Feature scaling for kernel Fisher discrim inant analysis using leave-one-out cross vali dation. FS-KFDA is a package for implementing f eature scaling for kernel fisher discriminant analysis.

clear
load blfsegment;
load segmentindex;
trainnum = 500;
starttime = cputime;
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
for k = 1:10
    % Tanslate the data
    index1            =  segmentindex（k1:trainnum）;
    index2            =  segmentindex（k1+trainnum:end）;
    trainsamples      =  segmentsamples（index1:）;
    testsamples       =  segmentsamples（index2:）;
    [trainsamplesAB]=  scaletrain（trainsamples）;
    testsamples       =  scaletest（testsamplesAB）;
    trainlabels       =  segmentlabels（index1）;
    testlabels        =  segmentlabels（index2）;
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%   calculate the correct rate       
      
    options = optimset（‘GradObj‘‘on‘）;
    options = optimset（options‘LargeScale‘‘off‘）;
    options = optimset（options‘DerivativeCheck‘‘off‘）;
    options = optimset（options‘Display‘‘iter‘）;
    options = optimset（options‘MaxIter‘50）;
    options = optimset（options‘TolFun‘1e-4）;
    options = optimset（options‘TolX‘1e-4）;
    options = optimset（options‘LineSearchType‘‘cubicpoly‘）;
    
    D = size（trainsamples2）;
    x = log（1/D）*ones（size（trainsamples2）1）;
    [xfxc]     = fminunc（‘mkfdakernel‘[x;0*ones（11）]optionstrainsamplestrainlabels）;
    [abloo（k）] = mkfdakernel（xtrainsamplestrainlabels）;
    fprintf（‘%f\n‘loo（k））;
    result       = mkfdapred（x trainsamples trainlabelstestsamples）;
    testerror（k） = 1 - mean（result == testlabels）;
     
     fprintf（‘meanerror = %f\n‘mean（testerror））;
     fprintf（‘Iteration = %d\n‘k）;
 end
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%   print the output
endtime   = cputime;
meanerror = mean（testerror）;
stderror  = std（testerror）;
fprintf（‘meanerror = %f\n‘meanerror）;
fprintf（‘stderror  = %f\n‘stderror）;
fprintf（‘time      = %f\n‘endtime-starttime）;
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

 属性            大小     日期    时间   名称
----------- ---------  ---------- -----  ----

     文件       2501  2006-10-10 20:18  fs-kfda 1.0\bench_segment.m

     文件     335224  2006-10-10 17:39  fs-kfda 1.0\blfsegment.mat

     文件        712  2006-10-10 20:26  fs-kfda 1.0\content.m

     文件        568  2006-01-16 11:05  fs-kfda 1.0\evalkernel.m

     文件       1749  2005-07-08 11:22  fs-kfda 1.0\kfdakernel.m

     文件       1186  2005-06-29 11:13  fs-kfda 1.0\kfdapred.m

     文件       2127  2005-04-08 20:18  fs-kfda 1.0\mkfdakernel.m

     文件       1525  2005-04-06 20:36  fs-kfda 1.0\mkfdapred.m

     文件        180  2005-01-21 20:18  fs-kfda 1.0\scaletest.m

     文件        231  2005-06-11 09:41  fs-kfda 1.0\scaletrain.m

     文件     380757  2006-10-10 19:22  fs-kfda 1.0\segmentindex.mat

     文件        380  2011-03-07 11:24  fs-kfda 1.0\注释.txt

     目录          0  2006-10-10 20:15  fs-kfda 1.0

----------- ---------  ---------- -----  ----

               727140                    13