Projects/Project2.R at master · iamSarang/Projects · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
#Clear the environment

rm(list=ls(all=TRUE))

#Read the input data that is given

setwd("D:/INSOFE/CSE7305c/CUTe 3")

company_data<-read.csv("train.csv",header = T)

test_data<-read.csv("test.csv",header=T)

#Use head() and tail() functions to get a feel of the data

head(company_data)

tail(company_data)

head(test_data)

tail(test_data)

#Check the structure of the input data

str(company_data)

str(test_data)

#Check the distribution of the input data using the summary function

summary(company_data)

summary(test_data)

sum(is.na(company_data))

sum(is.na(test_data))

#Removing columns that are not necessary
company_data_mod<-company_data[,-c(1,66)]

test_data_mod<-test_data[-c(1)]

#Impute and standardize the data

library(DMwR)
library(caret)

company_data[is.na(company_data)]<-0

company_data_final<-company_data

test_data[is.na(test_data)]<-0

test_data_std<-test_data

company_data_final$target<-as.factor(as.character(company_data_final$target))

# Divide the data into train and validation
set.seed(123)

train_RowIDs = createDataPartition(company_data_final$target,p=0.8,list=F)
train = company_data_final[train_RowIDs,]
validation= company_data_final[-train_RowIDs,]
rm(train_RowIDs)

#Exclude target variable from train/validation data before building the model
train_Data_wo_target <- train[,-which(names(train) %in% c("target"))]
validation_Data_wo_target <- validation[,-which(names(train) %in% c("target"))]

#Build ensemble model using adaboost
install.packages("ada")
library(ada)

model_ada = ada(x = train_Data_wo_target,
            y = train$target,
            iter=400, loss="exponential", type= "discrete", nu= 0.45)

pred_Train  =  predict(model_ada, train_Data_wo_target)

#Using the model built, predict the values for validation data
pred_Validation = predict(model_ada, validation_Data_wo_target)

#Check the accuracy on train/validation data
cm_Train = table(train$target, pred_Train)
accu_Train= sum(diag(cm_Train))/sum(cm_Train)
accu_Train

cm_Validation = table(validation$target, pred_Validation)
accu_Validation= sum(diag(cm_Validation))/sum(cm_Validation)
accu_Validation

#Print the confusion Matrix

conf_matrix_Train <- table(train$target, pred_Train)

conf_matrix_Train

recall_Train <- conf_matrix_Train[2, 2]/sum(conf_matrix_Train[2, ])

precision_Train<-conf_matrix_Train[2,2]/sum(conf_matrix_Train[,2])

F1_Train <- (2 * precision_Train * recall_Train) / (precision_Train + recall_Train)
F1_Train

conf_matrix_Validation <- table(validation$target, pred_Validation)

conf_matrix_Validation

recall_Validation <- conf_matrix_Validation[2, 2]/sum(conf_matrix_Validation[2, ])

precision_Validation<-conf_matrix_Validation[2,2]/sum(conf_matrix_Validation[,2])

F1_Validation <- (2 * precision_Validation * recall_Validation) / (precision_Validation + recall_Validation)
F1_Validation

#Using the model built, predict the values for test data
pred_Test  =  predict(model_ada, test_data_std)

plot(model_ada)

#Write the output file

output<-data.frame(test_data$ID,pred_Test)

summary(output)

colnames(output)<-c("ID","prediction")

write.csv(output,file="Samplesubmission.csv")