Cancer <- read.table("http://www.cse.chalmers.se/~chrdimi/downloads/fouille/cancer.txt")
## http://bit.ly/1OvaqCP

## Let us count the incidence of cancer in our dataset
nData <- dim(Cancer)[1]
nCancer <- sum(Cancer$cancer == 1)

pCancer <- nCancer / nData

print("Proportion of cancer patients")
print(pCancer)

## First, count how many males nad females there are
nMales <- sum(Cancer$sex == "male")
nFemales <- sum(Cancer$sex == "female")

## First, count how many males have cancer
maleCancer <- sum(Cancer[Cancer$sex=="male",]$cancer)

## Then, count how many females have cancer
femaleCancer <- sum(Cancer[Cancer$sex=="female",]$cancer)


print("Male cancer incidence")
print(pCancerMale)

print("Female cancer incidence")
print(pCancerFemale)

pCancerMale <- maleCancer / nMales
pCancerFemale <- femaleCancer / nFemales
pCancer <- (maleCancer + femaleCancer) / (nMales + nFemales)

print("Overall cancer incidence (same as before)")
print(pCancer)


## Select only the data pertaining to age and cancer
X = Cancer[c("age", "cancer")]

## example of how we can actually perhaps classify by age
pdf("agePlot1.pdf", width = 4, height = 4)
plot(X[1:10,])
dev.off()

pdf("agePlot2.pdf", width = 4, height = 4)
plot(X[1:20,])
dev.off()

## example of how we can actually perhaps classify by age
##plot(X[1:20,])