glass

glass<-read.csv("C:/Users/blacksnake/Downloads/glass.csv")
prop.table(table(glass$Type))

glass$Type<-factor(glass$Type,levels= c("1","2","3","5","6","7"),labels = c("building_windows_float_processed","building_windows_non_float_processed","vehicle_windows_float_processed","containers","tableware","headlamps"))
round(prop.table(table(glass$Type))*100,digits = 1)

normalize<-function(x){
  return((x-min(x))/(max(x)-min(x)))
}

glass_n<-as.data.frame(lapply(glass[1:9], normalize))
glass_n1<-cbind(glass_n,glass$Type)

glass_train<-glass_n[1:198,]
glass_test<-glass_n[199:214,]

glass_train_labels<-glass[1:198,10]
glass_test_labels<-glass[199:214,10]

#install.packages("class")
library(class)

#install.packages("gmodels")
library(gmodels)

test_acc <- NULL
train_acc <- NULL
for (i in seq(3,200,2))
{
  train_glass_pred <- knn(train=glass_train,test=glass_train,cl=glass_train_labels,k=i)
  train_acc <- c(train_acc,mean(train_glass_pred==glass_train_labels))
  test_glass_pred <- knn(train = glass_train, test = glass_test, cl = glass_train_labels, k=i)
  test_acc <- c(test_acc,mean(test_glass_pred==glass_test_labels))
}



par(mfrow=c(1,2)) # c(1,2) => indicates 1 row and 2 columns
plot(seq(3,200,2),train_acc,type="l",main="Train_accuracy",col="blue")
plot(seq(3,200,2),test_acc,type="l",main="Test_accuracy",col="red")

acc_neigh_df <- data.frame(list(train_acc=train_acc,test_acc=test_acc,neigh=seq(3,200,2)))

install.packages("ggplot2")
library(ggplot2)
ggplot(acc_neigh_df,aes(x=neigh))+
  geom_line(aes(y=train_acc,colour="train_acc"),lwd=1.5)+
  geom_line(aes(y=test_acc,colour="test_acc"),lwd=1.5)+
  scale_fill_manual(" ",breaks=c("train_acc","test_acc"),values = c("train_acc"="green","test_acc"="red"))


glass_pred <- knn(train = glass_train, test = glass_test, cl = glass_train_labels, k=14)