library(tree)

setwd("C:/Users/Alexandre/Downloads")
salaire<-as.data.frame(read.csv(file = "Salaries.csv", header = T, sep = ",", dec="."))[,-1]
summary(data)

tree.Lin <- tree( salary ~ yrs.service + yrs.since.phd , data=salaire)
plot( tree.Lin )
text( tree.Lin , cex=.75)

#a) phd > 32 ans
#b) 4

#c)
tree.model<-tree(log(salary) ~ yrs.service + yrs.since.phd , data=salaire)
plot( tree.model )
text( tree.model , cex=.75)

#d) les salaires suivent une loi Log-Normale. Il est donc normal de le passer au log


## 1.3

salar.deciles <- quantile( salaire$salary, 0:10/10)
cut.prices <- cut( salaire$salary , salar.deciles , include.lowest = TRUE )
plot( salaire$yrs.service , salaire$yrs.since.phd , col = grey(10:2/11)[cut.prices], pch =20 , xlab ="yrs.service" , ylab ="yrs.since . phd ")
partition.tree( tree.model , ordvars = c("yrs.service" ,"yrs.since.phd") , add = TRUE )
plot ( salaire$yrs.since.phd , salaire$salary , pch =19 , col = as.numeric (salaire$rank ) )
partition.tree( tree.model , label ="Species" , add = TRUE )
legend("topright" , legend = unique(salaire$rank) , col = unique ( as.numeric( salaire$rank ) ) , pch =19)


## 1.4

summary ( tree.model )
tree.model2 <- tree( log( salary ) ~ yrs.service + yrs.since.phd , data = salaire , mindev =0.001)
plot ( tree.model2 )
text ( tree.model2 , cex =.75)
summary ( tree.model2 )
pruned.tree <- prune.tree( tree.model , best =4)
plot(pruned.tree)
text(pruned.tree)

### II