%Jake knife test code
clc;
close all;
clear all;
addpath libsvm-mat-2.88-1;
load iOri_NAC_TriNAC_PseDNC_features;
%****************************************
Result=0;
y=[];
Accuracy=0;
Total_Seq_train=811;
DNA_labels=[];
Total_correct=0;
c1=0;
c2=0;
DNA_labels(1:405)=-1;
DNA_labels(406:811)=1;
%+++++++++++++++++++++++++++++ train label
Labelstem=[];
Samplestem=[];
Samplestem=iOri_NAC_TriNAC_PseDNC_features;
Labelstem= DNA_labels';
Knn=2;
for A=1:size(Samplestem,1)
A
if A==1
Samples=Samplestem(A+1:end,:)';
TestSample=Samplestem(A,:)';
Labels=Labelstem(A+1:end,:)';
TestLabel=Labelstem(A,:)';
else
s11=Samplestem(1:(A-1),: ); % Jackknifing
s22=Samplestem((A+1):end,:);
Samples=[s11;s22]';
TestSample=Samplestem(A,:)';
l11=Labelstem(1:(A-1),: );
l22=Labelstem((A+1):end,:);
Labels=[l11;l22]';
TestLabel=Labelstem(A,:)';
end
%ResultLabel = Nearest_Neighbor(Samples, Labels, TestSample, Knn);
% y = [y ResultLabel];
model = svmtrain (Labels' , Samples' , '-t 2 -c 1.5 -g 0.0005');
[Predict_label,accuracy, dec_values] = svmpredict(TestLabel, TestSample', model);
% Mahlabadistane_PNN_2_kPCA_dipep(A,Predict_label )=1; % 1=true for the class which has won
y(A)=Predict_label;
end
y2=y;
Result=find(y==DNA_labels);
Total_correct=size(Result,2);
Accuracy=(Total_correct/Total_Seq_train)*100
%+++++++++ individual Accuracy
for i=1:405
if( y(i)==-1)
c1=c1+1;
end
end
for i=406:811
if( y(i)==1)
c2=c2+1;
end
end
C1=(c1/405)*100
C2=(c2/406)*100
fwd.zip_The Dos
版权申诉
116 浏览量
2022-09-24
07:51:18
上传
评论
收藏 225KB ZIP 举报
weixin_42651887
- 粉丝: 79
- 资源: 1万+