R code for classification on the UC Irvine Wine data

##Uncomment it the "glmnet" package is never previously installed

##install.packages("glmnet");

library(glmnet);

tmp<-read.csv("winequality.csv", header=TRUE, sep=";");

n<-nrow(tmp);

K<-ncol(tmp)-1;

x<-matrix(0,n,K);

for(i in 1:K) {x[,i]<-tmp[,i];}

y<-tmp[,K+1];

##To split the data into training and test set

idx<-sample(1:n,floor(n/2),replace=FALSE);

xtr<-x[idx,]; ytr<-y[idx];

xts<-x[-idx,]; yts<-y[-idx];

##Fit the multiple logistic regression model

myglm4<-glmnet(xtr,ytr,family="multinomial");

##Apply the trained model to the test set

mypred4<-predict(myglm4,newx=xts,type="response",s=0.01);

posteriprob<-mypred4[,,1];

yhat<-matrix(1,nrow(xts),1);

for(i in 1:nrow(xts))

{

yhat[i]<-which.max(posteriprob[i,]);

}

acc<-sum(yhat+2==yts)/nrow(xts);

cat("Accuracy on the test set is", acc, "\n");