yellow-bike

install.packages("readr")
library(readr)

setwd("E:")
train<-read_csv("train.csv")
test<-read_csv("test.csv")
head(data)

test$registered=0
test$casual=0
test$count=0
data<-rbind(train,test)

install.packages("lubridate")
library(lubridate)
install.packages("randomForest")
library(randomForest)
install.packages("rpart")
install.packages("rpart.plot")
library(rpart)
library(rpart.plot)

#data$hour1<-hour(data$datetime)
data$hour1<-substr(data$datetime,12,13)
table(data$hour1)

data$hour1<-as.integer(data$hour1)
train<-data[as.integer(day(data$datetime))<20,]
boxplot(train$count~train$hour1,xlab="hour",ylab="count")

boxplot(train$registered~train$hour1,xlab="hour",ylab="registered count")
boxplot(train$casual~train$hour1,xlab="hour",ylab="casual count")

data$day<-weekdays(as.Date(data$datetime))
table(data$day)
data$year<-year(as.Date(data$datetime))
data$month<-month(as.Date(data$datetime))

train<-data[as.integer(day(data$datetime))<20,]
boxplot(train$count~train$day,xlab="day",ylab="count")
boxplot(train$registered~train$day,xlab="day",ylab="registered count")
boxplot(train$casual~train$day,xlab="day",ylab="casual count")

train<-data[as.integer(day(data$datetime))<20,]
boxplot(train$count~train$weather,xlab="year",ylab="count")
boxplot(train$registered~train$day,xlab="weather",ylab="registered count")
boxplot(train$casual~train$day,xlab="weather",ylab="casual count")

cor_data<-data.frame(train$temp,train$atemp,train$humidity,train$windspeed,train$casual,train$registered)
cor(cor_data)

data$year<-year(data$datetime)
train<-data[as.integer(day(data$datetime))<20,]
boxplot(train$count~train$year,xlab="year",ylab="count")

########

train$hour1<-as.integer(train$hour1)
d<-rpart(registered~hour1,data=train)
rpart.plot(d)

train$hour1<-as.integer(train$hour1)
data$dp_reg=0
data$dp_reg[data$hour1<7.5]=1
data$dp_reg[data$hour1>=22]=2
data$dp_reg[data$hour1>=9.5 & data$hour1<18]=3
data$dp_reg[data$hour1>=7.5 & data$hour1<8.5]=4
data$dp_reg[data$hour1>=8.5 & data$hour1<9.5]=5
data$dp_reg[data$hour1>=20 & data$hour1<22]=6
data$dp_reg[data$hour1>=18 & data$hour1<20]=7

data$dp_cas=0
f=rpart(casual~hour1,data=train)
rpart.plot(f)

data$dp_cas[data$hour1<7.5]=1
data$dp_cas[data$hour1>=20]=2
data$dp_cas[data$hour1>7.5 & data$hour1<9.5]=3
data$dp_cas[data$hour1>=9.5 & data$hour1<20]=4

g<-rpart(registered~temp,data=train)
rpart.plot(g)

data$temp_reg=0
data$temp_reg[data$temp<13]=1
data$temp_reg[data$temp>=30]=2
data$temp_reg[data$temp>=13 & data$temp<23]=3
data$temp_reg[data$temp>=23 & data$temp<30]=4

data$temp_cas=0
h<-rpart(casual~temp,data=train)
rpart.plot(h)

data$temp_cas=0
data$temp_cas[data$temp<15]=1
data$temp_cas[data$temp>=30]=2
data$temp_cas[data$temp>=15 & data$temp<23] = 3
data$temp_cas[data$temp>=23 & data$temp<30] = 4

data$year_part=0
data$month<-month(data$datetime)
data$year_part[data$year=='2011']=1
data$year_part[data$year=='2011' & data$month>3]=2
data$year_part[data$year=='2011' & data$month>6]=3
data$year_part[data$year=='2011' & data$month>9]=4
data$year_part[data$year=='2012']=5
data$year_part[data$year=='2012' & data$month>3]=6
data$year_part[data$year=='2012' & data$month>6]=7
data$year_part[data$year=='2012' & data$month>9]=8

table(data$year_part)

data$day_type=""
data$day_type[data$holiday==0 & data$workingday==0]="weekend"
data$day_type[data$holiday==1]="holiday"
data$day_type[data$holiday==0 & data$workingday==1]="working day"
data$weekend=0
data$weekend[data$day=="Sunday"|data$day=="Saturday"]=1

#####randomForest####

#train<-data[as.integer(day(data$datetime))>20,]
train<-data

set.seed(421)
train$logreg<-log(train$registered+1)
train$logcas<-log(train$casual+1)

#fit1<-randomForest(logreg~hour1+workingday+day+holiday+day_type+
                   temp_reg+humidity+atemp+windspeed+season+weather
                   +dp_reg+weekend+year+year_part,
                   train,importance=TRUE,ntree=250)

fit1<-randomForest(logreg~hour1+workingday+holiday+temp_reg+atemp+humidity
                    +windspeed+season+weather+weekend+year+year_part
                    +dp_reg,train,importance=TRUE,ntree=500)

pred1<-predict(fit1,train)
train$logreg<-pred1

set.seed(421)
fit2<-randomForest(logcas~hour1+workingday+holiday+temp_cas+atemp+humidity
                   +windspeed+season+weather+weekend+year+year_part
                   +dp_cas,train,importance=TRUE,ntree=500)

pred2<-predict(fit2,train)
train$logcas<-pred2

train$registered<-exp(train$logreg)-1
train$casual<-exp(train$logcas)-1
train$count<-train$casual+train$registered

train2<-train[as.integer(day(data$datetime))>=20,]

submit_final<-data.frame(datetime=train2$datetime,count=train2$count)
write.csv(submit_final,"submit_final.csv",row.names=F)