finding classifier performance in MATLAB

3 次查看（过去 30 天）

Ramya Mohan 2013-2-4

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/62186-finding-classifier-performance-in-matlab

编辑： Walter Roberson 2016-11-14

I want to classify the breast cancer dataset using SVM.

I divide the 569 instances into 5 sets of data and groups, then use classperf function to compute the performance, but I don't know what arguments are to be given in classperf function. Pls help me.

============================================================

clc;clf;clear all;close all;
data=xlsread('DATACANCER.XLSX');
load clasy;
 i=24;
 j=25;
data = [data(:,i), data(:,j)];
A1= data(1:113,:);A2=data(114:226,:);A3=data(227:340,:);A4=data(341:454,:);A5=data(455:569,:);
groups = ismember(species,'B');
[train, test] = crossvalind('holdOut',groups,0.2);
G1=groups(1:113,:);G2=groups(114:226,:);G3=groups(227:340,:);G4=groups(341:454,:);G5=groups(455:569,:);
cp = classperf(groups);
figure
svmStruct = svmtrain(A1,G1,'showplot',true);
title(sprintf('Kernel Function: %s',...
              func2str(svmStruct.KernelFunction)),...
              'interpreter','none');
         classes = svmclassify(svmStruct,A2,'showplot',true);
         classperf(cp,classes,----);<--------HERE IS MY DOUBT
         ans1=cp.CorrectRate*100;
disp(ans1);
cp
        figure
svmStruct = svmtrain(A2,G2,'showplot',true);
title(sprintf('Kernel Function: %s',...
              func2str(svmStruct.KernelFunction)),...
              'interpreter','none');
         classes = svmclassify(svmStruct,A3,'showplot',true);
         classperf(cp,classes,----);<--------HERE IS MY DOUBT
ans2=cp.CorrectRate*100;
disp(ans2);
cp
 figure
svmStruct = svmtrain(A3,G3,'showplot',true);
title(sprintf('Kernel Function: %s',...
              func2str(svmStruct.KernelFunction)),...
              'interpreter','none');
         classes = svmclassify(svmStruct,A4,'showplot',true);
         classperf(cp,classes,--------);<--------HERE IS MY DOUBT
ans3=cp.CorrectRate*100;
disp(ans3);
cp
   figure
  svmStruct = svmtrain(A4,G4,'showplot',true);
  title(sprintf('Kernel Function: %s',...
                func2str(svmStruct.KernelFunction)),...
                'interpreter','none');
           classes = svmclassify(svmStruct,A5,'showplot',true);
           classperf(cp,classes,-------);<--------HERE IS MY DOUBT
  ans4=cp.CorrectRate*100;
  disp(ans4);
  cp
   figure
  svmStruct = svmtrain(A5,G5,'showplot',true);
  title(sprintf('Kernel Function: %s',...
                func2str(svmStruct.KernelFunction)),...
                'interpreter','none');
           classes = svmclassify(svmStruct,A1,'showplot',true);
           classperf(cp,classes,-------);<--------HERE IS MY DOUBT
  ans5=cp.CorrectRate*100;
  disp(ans5);
  cp