library(MASS)
painters
painters_sg<-painters[, 1:4]

print("Skladowe glowne: ")
sklad_glowne=prcomp(painters_sg)
sklad_glowne
plot(sklad_glowne)
summary(sklad_glowne)
biplot(sklad_glowne)

print("Wartosci srednie: ")

print("Composition: ")
mean(painters[,1])

print("Drawings: ")
mean(painters[,2])

print("Colors: ")
mean(painters[,3])

print("Expressions: ")
mean(painters[,4])

print("Odchylenia standardowe: ")

print("Composition: ")
sd(painters[,1])

print("Drawings: ")
sd(painters[,2])

print("Color: ")
sd(painters[,3])

print("Expression: ")
sd(painters[,4])

print("Histogram: ")
summary(painters)

print("Composition: ")
hist_comp<-subset(painters, painters$Composition<20)
summary(hist_comp$Composition)
attach(hist_comp)
summary(Composition)
bin_comp=seq(min(Composition),max(Composition)+2,2)
hist(Composition, main="Histogram composition", ylab="Czestotl composition", xlab="Composition", col="grey", breaks=54)

print("Drawing: ")
hist_draw<-subset(painters, painters$Drawing<20)
summary(hist_draw$Drawing)
attach(hist_draw)
summary(Drawing)
bin_draw=seq(min(Drawing),max(Drawing)+2,2)
hist(Drawing, main="Histogram drawing", ylab="Czestotl drawing", xlab="Drawing", col="blue", breaks=54)

print("Color: ")
hist_col<-subset(painters, painters$Colour<20)
summary(hist_col$Colour)
attach(hist_col)
summary(Colour)
bin_col=seq(min(Colour),max(Colour)+2,2)
hist(Colour, main="Histogram color", ylab="Czestotl color", xlab="Colour", col="red", breaks=54)

print("Expression: ")
hist_exp<-subset(painters, painters$Expression<20)
summary(hist_exp$Expression)
attach(hist_exp)
summary(Expression)
bin_exp=seq(min(Expression),max(Expression)+2,2)
hist(Expression, main="Histogram of 'Expression'", ylab="Czestotl expression", xlab="Expression", col="yellow", breaks=54)

print("Macierz korelacji: ")
pairs(painters[,1:4])
cor(painters[,1:4])
cor(painters[,1:4], method="spearman")
cor(painters[,1:4], method="kendall")

print("Klasyfikator LDA")
painters.lda=lda(School~., data=painters)
painters.pred=predict(painters.lda, newdata=painters)

print(table_lda<-table(painters$School, painters.pred$class))

print(procent<-100*sum(diag(table_lda))/sum(table_lda))

print("Klasyfikator QDA")
painters.qda1=qda(School~Composition+Drawing, data=painters)
painters.pred_qda1=predict(painters.qda1, newdata=painters)

print(table_qda1<-table(painters$School, painters.pred_qda1$class))

print(procent<-100*sum(diag(table_qda1))/sum(table_qda1))
painters.qda2=qda(School~Colour+Expression, data=painters)
painters.pred_qda2=predict(painters.qda2, newdata=painters)

print(table_qda2<-table(painters$School, painters.pred_qda2$class))

print(procent<-100*sum(diag(table_qda2))/sum(table_qda2))

print("Metoda krokowa")
set.seed(4578)
painters.step_forw=stepclass(School~., data=painters, method="lda", direction="forward", improvement=0.0001)
painters.step_back=stepclass(School~., data=painters, method="lda", direction="backward", improvement=0.0001)

print("Metoda Bayesa")
painters.bayes<-NaiveBayes(School~., data=painters, userkernel=TRUE)
painters.bayes_pred<-predict(painters.bayes, painters)

print(table_bayes<-table(painters$School, painters.bayes_pred$class))

print(procent<-100.0*sum(diag(table_bayes))/sum(table_bayes))

print("Metoda najblizszych sasiadow")
data(painters)
str(painters)
table(painters$School)
head(painters)
set.seed(9850)
painters.grp<-runif(nrow(painters))
painters<-painters[order(painters.grp),]
str(painters)
summary(painters[,c(1,2,3,4)])
normalizacja<-function(x){ return( (x-min(x))/(max(x)-min(x)) ) }
paintersNorm<-as.data.frame(lapply(painters[,c(1,2,3,4)], normalizacja))
str(paintersNorm)
summary(paintersNorm)
paintersTrn<-paintersNorm[1:33,]
painters_test<-paintersNorm[34:54,]
paintersTrn_cel<-painters[1:33,5]
painters_test_cel<-painters[34:54,5]
require(class)
model1<-knn(train=paintersTrn, test=painters_test, cl=paintersTrn_cel, k=8)
model1
knnTable=table(painters_test_cel, model1)
knnTable

print(procent<-100*sum(diag(knnTable))/sum(knnTable))