Scribd
Upload
31 views5 pages

Code Documentation: Loading The Packages

This document outlines code for analyzing pulsar star data. It loads packages, reads in training and test data, joins the datasets and renames columns. It checks for NAs, plots correlations, performs imputation, scales features, splits the data into training, validation and test sets. A support vector machine (SVM) model with a linear kernel is trained on the training set and validated on the held-out validation set. Predictions are made on the test set and a confusion matrix is plotted to evaluate model performance. The predicted test labels are written to a CSV file.

Uploaded by

Shashwat Patel mm19b053
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
31 views5 pages

Code Documentation: Loading The Packages

This document outlines code for analyzing pulsar star data. It loads packages, reads in training and test data, joins the datasets and renames columns. It checks for NAs, plots correlations, performs imputation, scales features, splits the data into training, validation and test sets. A support vector machine (SVM) model with a linear kernel is trained on the training set and validated on the held-out validation set. Predictions are made on the test set and a confusion matrix is plotted to evaluate model performance. The predicted test labels are written to a CSV file.

Uploaded by

Shashwat Patel mm19b053
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

Code Documentation

Loading the packages

library(tidyverse)
library(mice)
library(ggcorrplot)
library(ggthemes)
library(e1071)
library(caret)
library(scales)

Reading the data

train<-read_csv("pulsar_data_train.csv")
test<-read_csv("pulsar_data_test.csv")

Renaming the columns of the dataset and joining both test and
train data

colnames(train)<-c("Mean_IP","SD_IP","EK_IP","Skew_IP","Mean_DMSNR","SD_DMSNR",
"EK_DMSNR","Skew_DMSNR","class")

colnames(test)<-c("Mean_IP","SD_IP","EK_IP","Skew_IP","Mean_DMSNR","SD_DMSNR",
"EK_DMSNR","Skew_DMSNR","class")

full_data<-full_join(train,test)

Function to check number of NAs in the data

na_check<-function(dataset){
sapply(dataset,function(x) sum(is.na(x)))
}

1
Basic summary of data

str(full_data)
summary(full_data)

na_check(full_data[,-9])

na_check(train)

na_check(test[,-9])

Correlation plot

correlation<-cor(full_data[,-9],use="na.or.complete")

ggcorrplot(correlation, hc.order = TRUE,lab = TRUE)

Imputation and data preprocessing

class<-full_data[,9]
full_dat_without_class<-full_data[-9]

imputed_Data <- mice(full_dat_without_class, m=5, maxit = 50, method = 'pmm', seed = 500)
completeData <- complete(imputed_Data,2)

full_data<-completeData
full_data$class<-class$class

full_data$class<-as.factor(full_data$class)

Setting up the theme for visualization

my_theme<-theme_fivethirtyeight()+theme(plot.title = element_text(hjust = 0.5,size=20),


axis.title = element_text(size=20),
axis.text = element_text(size=14),
plot.subtitle = element_text(hjust=0.5),
legend.position = "top",
legend.title = element_text(size=15),
legend.text = element_text(size=15))

Fig 4

2
full_data%>%filter(!is.na(class))%>%ggplot()+
geom_point(aes(Mean_IP,Skew_IP,col=class),size=2)+
scale_color_brewer(palette = "Set1")+
labs(title="Mean Vs Skewness of Integrated Profile")+
my_theme

Fig 5

full_data%>%filter(!is.na(class))%>%ggplot()+
geom_point(aes(Skew_IP,EK_IP,col=class),size=2)+
scale_color_brewer(palette = "Set1")+
labs(title="Skewness Vs Excess kurtosis of Integrated Profile")+
my_theme

Fig 6

full_data%>%filter(!is.na(class))%>%ggplot()+
geom_point(aes(Mean_IP,SD_IP,col=class),size=2)+
scale_color_brewer(palette = "Set1")+
labs(title="Mean Vs Standard Deviation of Integrated Profile")+
my_theme

Fig 7

full_data%>%filter(!is.na(class))%>%ggplot()+
geom_point(aes(Mean_DMSNR,SD_DMSNR,col=class),size=2)+
scale_color_brewer(palette = "Set1")+
labs(title="Mean Vs Standard Deviation of DM-SNR curve")+
my_theme

Fig 8

full_data%>%filter(!is.na(class))%>%ggplot()+
geom_point(aes(Mean_DMSNR,EK_DMSNR,col=class),size=2)+
scale_color_brewer(palette = "Set1")+
labs(title="Mean Vs Excess kurtosis of DM-SNR curve")+
my_theme

3
Fig 9

full_data%>%filter(!is.na(class))%>%ggplot()+
geom_point(aes(Skew_DMSNR,EK_DMSNR,col=class),size=2)+
scale_color_brewer(palette = "Set1")+
labs(title="Skewness Vs Excess kurtosis of DM-SNR curve")+
my_theme

Feature Scaling

full_data[,1:8]<-scale(full_data[,1:8])

Splitting the training, validation, test dataset

set.seed(1)

train<-full_data%>%filter(!is.na(class))

Index <- createDataPartition(train$class,p=0.8,list=FALSE)

train_data<-train[Index,]

validation_data<-train[-(Index),]

test<-full_data%>%filter(is.na(class))

Model

classifier_1<- svm(formula = class ~ .,


data = train_data,
type = 'C-classification',
kernel = 'linear')

val_pred<-predict(classifier_1,newdata = validation_data[,-9])

test_pred<- predict(classifier_1, newdata = test[,-9])

test$class<-test_pred

Confusion Matrix and Plot

4
cm_model1<-confusionMatrix(val_pred,validation_data$class)

cm_dataframe<-as.data.frame(cm_model1$table)

ggplot(data =cm_dataframe ,
aes(x = Reference, y = Prediction)) +
geom_tile(aes(fill = log(Freq)), colour = "white") +
scale_fill_gradient(low = "white", high = "steelblue") +
geom_text(aes(x = Reference, y = Prediction, label = Freq),size=6) +
labs(x="True Class",y="Predicted Class")+
ggtitle(paste("Accuracy:",percent_format()(cm_model1$overall[1])))+
theme_fivethirtyeight()+
theme(legend.position = "none",
axis.title = element_text(size=20),
axis.text = element_text(size=14),
plot.title = element_text(hjust=0.5,size=20))

Write predicted data csv

write.csv(test,file = "predicted.csv",row.names = FALSE)

You might also like