SVM&nnet使用代码笔记

results<-data.frame(matrix(0,ncol=96*2,nrow=length(testseq)))
  for(col in 1:96)
  {

    train<-data_reshape[[col]][trainseq,]
    test<-data_reshape[[col]][testseq,]
    model<-svm(X7~.,train)
    pre<-predict(model,subset(test,select=-X7))
    max<-data_param[col,1]
    min<-data_param[col,2]
    testOrigin<-(test[,7]*(max-min))+min
    preOrigin<-(pre*(max-min))+min
    results[,col]<-testOrigin
    results[,col+96]<-preOrigin
  }

  results<-data.frame(matrix(0,ncol=96*2,nrow=length(testseq)))
  for(col in 1:96)
  {
    train<-data_reshape[[col]][trainseq,]
    test<-data_reshape[[col]][testseq,]
    model<-nnet(X7~.,train,size=20,decay=0.01,maxit=600)
    pre<-predict(model,subset(test,select=-X7))
    max<-data_param[col,1]
    min<-data_param[col,2]
    testOrigin<-(test[,7]*(max-min))+min
    preOrigin<-(pre*(max-min))+min
    results[,col]<-testOrigin
    results[,col+96]<-preOrigin
  }