Normal view MARC view ISBD view

Increase the efficiency of loan prediction using machine learning approach / Nik Nur Ain Binti Nik Jid

By:

Nik Nur Ain Binti Nik Jid [author.]

Material type: Text

text

Media type:

unmediated

Carrier type:

volume

ISBN:

THE0010021 (Local)

Subject(s):

Dissertation note: Bachelor of Applied Science in Data Analytics with Honours -- Universiti Malaysia Pahang – 2023 Abstract: The individual loan approval process is a critical and ongoing issue faced by financial institutions. As the number of loan applications increases year by year, it is crucial for bankers to take the approval process seriously to prevent loan default cases. The goal of this study is to find the best prediction model that can assist financial institutions in overcoming this problem. Loan approval prediction involves determining the likelihood of a loan application being approved or denied by a financial institution. This typically involves analysing the applicant's credit history, income, employment status, and other financial factors. The objective is to reduce risk for the lender and increase the chances of loan approval for the borrower. In this study, machine learning techniques, such as logistic regression, decision trees, support vector machines, random forest, gradient booster, and artificial neural networks, were employed to train models for loan approval prediction using a home loan dataset obtained from Kaggle. The dataset includes variables such as gender, marital status, education, number of dependents, income, loan amount, and credit history. The analysis and modelling were conducted using Python in Google Colaboratory, also used as visualization tools aiding in identifying patterns and trends in the data. The results of the project revealed that the decision trees model exhibited the highest accuracy of 94% and the lowest root mean squared error value of 0.25, making it the most effective model for loan approval prediction. To deploy the prediction model, a web application was developed using Streamlit and Visual Studio Code as the integrated development environment platform. The web application provided an interactive graphical user interface and dashboard for end-users, enabling financial institutions to easily input loan application information and receive predictions on the likelihood of loan approval. By utilizing this prediction model, financial institutions can make more informed lending decisions, ultimately reducing the risk of loan default cases.

Tags from this library: No tags from this library for this title. Log in to add tags.

Average rating: 0.0 (0 votes)

Holdings ( 1 )
Title notes ( 4 )

Holdings
Item type	Current library	Call number	Status	Date due	Barcode
Final Year Report	UMPLIB GAMBANG	CD13576 (Browse shelf(Opens below))	In Transit		T000003174

Centre for Mathematical Sciences

Bachelor of Applied Science in Data Analytics with Honours -- Universiti Malaysia Pahang – 2023

Includes bibliographical references

The individual loan approval process is a critical and ongoing issue faced by financial institutions. As the number of loan applications increases year by year, it is crucial for bankers to take the approval process seriously to prevent loan default cases. The goal of this study is to find the best prediction model that can assist financial institutions in overcoming this problem. Loan approval prediction involves determining the likelihood of a loan application being approved or denied by a financial institution. This typically involves analysing the applicant's credit history, income, employment status, and other financial factors. The objective is to reduce risk for the lender and increase the chances of loan approval for the borrower. In this study, machine learning techniques, such as logistic regression, decision trees, support vector machines, random forest, gradient booster, and artificial neural networks, were employed to train models for loan approval prediction using a home loan dataset obtained from Kaggle. The dataset includes variables such as gender, marital status, education, number of dependents, income, loan amount, and credit history. The analysis and modelling were conducted using Python in Google Colaboratory, also used as visualization tools aiding in identifying patterns and trends in the data. The results of the project revealed that the decision trees model exhibited the highest accuracy of 94% and the lowest root mean squared error value of 0.25, making it the most effective model for loan approval prediction. To deploy the prediction model, a web application was developed using Streamlit and Visual Studio Code as the integrated development environment platform. The web application provided an interactive graphical user interface and dashboard for end-users, enabling financial institutions to easily input loan application information and receive predictions on the likelihood of loan approval. By utilizing this prediction model, financial institutions can make more informed lending decisions, ultimately reducing the risk of loan default cases.