R code
Download as TXT, PDF0 likes27 views
This document loads libraries and data to perform predictive modeling and feature selection. It loads data, combines test and train sets, selects predictive features, preprocesses data with PCA, splits into train, validation and test sets, builds a random forest model on the train set and predicts on the validation and test sets to evaluate performance.
![rm(list=ls()); gc()
library(dplyr)
library(C50)
library(caret)
library(randomForest)
library(funModeling)
setwd("C:/Users/Manav/Documents/FORE/Term 5/Big Data/End_Term_Project/")
train <- read.csv("train.csv")
test <- read.csv("test.csv")
test$TARGET <- 0
to <- rbind(train, test)
to$TARGET <- as.factor(to$TARGET)
df <- data.frame(to)
xyz <- df_status(df)
fun_1 = xyz%>%filter(p_zeros<100)
fun_2=subset(fun_1, select = c(variable))
fun_3=subset(to, select = c(fun_2$variable))
fun_4 = subset(fun_3, select = -c(ID,TARGET))
pre <- preProcess(fun_4, method = "pca")
inm <- predict(pre,fun_4)
inm$ID <- fun_3[,c("ID")]
inm$TARGET <- fun_3[,c("TARGET")]
inm[1:nrow(train),] -> tr
inm[-c(1:nrow(train)),] -> te
train_data <- tr[1:60000,]
valid_data <- tr[-c(1:60000),]
valid_data$TARGET <- NULL
te$TARGET <- NULL
model=randomForest(TARGET~.,data=train_data, ntree=100)
model
pred <- predict(model, valid_data)
confusionMatrix(pred, tr[-c(1:60000),c("TARGET")])
pred <- predict(model, te)](https://image.slidesharecdn.com/rcode-170120163218/85/R-code-1-320.jpg)
R code
- 1. rm(list=ls()); gc() library(dplyr) library(C50) library(caret) library(randomForest) library(funModeling) setwd("C:/Users/Manav/Documents/FORE/Term 5/Big Data/End_Term_Project/") train <- read.csv("train.csv") test <- read.csv("test.csv") test$TARGET <- 0 to <- rbind(train, test) to$TARGET <- as.factor(to$TARGET) df <- data.frame(to) xyz <- df_status(df) fun_1 = xyz%>%filter(p_zeros<100) fun_2=subset(fun_1, select = c(variable)) fun_3=subset(to, select = c(fun_2$variable)) fun_4 = subset(fun_3, select = -c(ID,TARGET)) pre <- preProcess(fun_4, method = "pca") inm <- predict(pre,fun_4) inm$ID <- fun_3[,c("ID")] inm$TARGET <- fun_3[,c("TARGET")] inm[1:nrow(train),] -> tr inm[-c(1:nrow(train)),] -> te train_data <- tr[1:60000,] valid_data <- tr[-c(1:60000),] valid_data$TARGET <- NULL te$TARGET <- NULL model=randomForest(TARGET~.,data=train_data, ntree=100) model pred <- predict(model, valid_data) confusionMatrix(pred, tr[-c(1:60000),c("TARGET")]) pred <- predict(model, te)