SVM算法训练的分类器

clc; clear all; close all; %% SVM分类器 n=2;%选择前面n个特征 Traindata=importdata('train.data'); Testdata=importdata('test.data'); Train=Traindata(:,1:10); %% 随机上采样 up=Traindata(randint(1,7819-466,[1 466]),:); Traindata_up=[Traindata(1:466,:);up;Traindata(467:8285,:)]; %% Traindataclass_up=Traindata_up(:,11); Trainclass=Traindata(:,11); Train2(find(Trainclass==-1))=-1; Train1(find(Trainclass==1))=1; Test=Testdata(:,1:10); Testclass=Testdata(:,11); Test2(find(Testclass==-1))=-1; Test1(find(Testclass==1))=1; %% training SVMStruct = svmtrain(Traindata_up(:,1:n), Traindataclass_up,'method', 'SMO','kernelcachelimit',50000); %% testing %Group = svmclassify(SVMStruct, Test(:,1:n)); Group = svmclassify(SVMStruct, Train(:,1:n)); %验证正确率（总） rightnumber=0; for i=1:size(Group,1) if Group(i)==Trainclass(i) %if Group(i)==Testclass(i) rightnumber=rightnumber+1; end end rightratio=rightnumber/size(Group,1)*100; sprintf('验证准确率（总）=%0.2f',rightratio) %验证正类正确率（1） rightnumber=0; for i=1:size(Train1,2) if Group(i)==Train1(i) rightnumber=rightnumber+1; end end rightratio=rightnumber/size(Train1,2)*100; sprintf('验证正类准确率=%0.2f',rightratio) %验证负类正确率（1） rightnumber=0; for i=size(Train1,2)+1:size(Train2,2) if Group(i)==Train2(i) rightnumber=rightnumber+1; end end rightratio=rightnumber/size(Train2,2)*100; sprintf('验证负类准确率=%0.2f',rightratio) %% %测试正确率（总） Group = svmclassify(SVMStruct, Test(:,1:n)); rightnumber=0; for i=1:size(Group,1) if Group(i)==Testclass(i) rightnumber=rightnumber+1; end end rightratio=rightnumber/size(Group,1)*100; sprintf('测试准确率（总）=%0.2f',rightratio) %测试正类正确率（1） rightnumber=0; for i=1:size(Test1,2) if Group(i)==Test1(i) rightnumber=rightnumber+1; end end rightratio=rightnumber/size(Test1,2)*100; sprintf('测试正类准确率=%0.2f',rightratio) %测试负类正确率（1） rightnumber=0; for i=size(Test1,2)+1:size(Test2,2) if Group(i)==Test2(i) rightnumber=rightnumber+1; end end rightratio=rightnumber/size(Test2,2)*100; sprintf('测试负类准确率=%0.2f',rightratio)