run_remove_noise6.rar_NOISE_reduce noise matlab

clear all close all clc load liver.mat a= liver; [n,m]=size(a); l=10*n/100; t=(1:345); l=round(l); r=randsample(t,l); test = a(r,:); tt=setdiff(t,r); train= a(tt,:); xtest=test(:,1:m-1); ytest=test(:,m); xtrain=train(:,1:m-1); ytrain=train(:,m); inc_node=1; sr=randomforest(xtest , xtrain , ytrain); s=svm1(xtest,xtrain,ytrain); C4 = C4_53(xtest, xtrain,ytrain,inc_node); %s=NaiveBayes.fit(xtrain , ytrain); NBModel = NaiveBayes.fit(xtrain,ytrain, 'Distribution', 'kernel'); Pred = NBModel.predict(xtest); sn=Pred; %Mdl=fitcknn(xtrain,ytrain,'NumNeighbors',4); knnModel=fitcknn(xtrain,ytrain,'NumNeighbors',4); pred=knnModel.predict(xtest); % rloss=resubLoss(Mdl); % CVMdl=crossval(Mdl); % kloss=kfoldLoss(CVMdl); sk=pred; FN=0;TN=0;FP=0; TP=0; NOISE=0; MAJORITY=0; i=1:l; o=1;ss=[]; noisee=[]; for i=1:l u=s(i)+sk(i)+sn(i)+sr(i); if u>0 maj(i)=1; else maj(i)=-1; end if (maj(i)==1 && ytest(i)==1) TP=TP+1; end if(maj(i)==-1 && ytest(i)==1) FP=FP+1; end if (maj(i)==1 &&ytest(i)==-1) FN=FN+1; end if (maj(i)==-1 &&ytest(i)==-1) TN=TN+1; end if (maj(i)==-1) ss=r(i); end disp(ss); noisee=union(noisee,ss); end DIF=setdiff(a,noisee); disp(DIF) recall =TP/(TP+FN); precision =TP/(TP+FP) ; accuracy = (TP+TN)/(TP+TN+FP+FN) ; recall=recall*100; precision=precision*100; accuracy=accuracy*100; disp(recall); disp(precision); disp(accuracy);