PREDICTION OF CUSTOMER CHURN FOR ABC MULTISTATE BANK USING MACHINE LEARNING ALGORITHMS - Abstract - Study Purpose and Key Results

Aug 4, 2026

Upload by Liam

Customer churn is the likelihood that customers stop doing business with a company within a defined period. ABC Multistate Bank faces the need to retain clients, making churn prediction central to more effective marketing and service planning. This study applies six supervised machine learning algorithms—K-Nearest Neighbors, Support Vector Machine, Naïve Bayes, Decision Tree, Random Forest, and XGBoost—using bank customer data from Kaggle. XGBoost achieves the best performance, enabling identification of at-risk customers for focused retention efforts and informing future model development.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages18

Reading Time45 min

AccessFree to download

Abstract

Introduction

Customer churn and its business impact

Data mining and machine learning background

Study contribution and scope

Literature Review

Review overview

PREDICTION OF CUSTOMER CHURN FOR ABC MULTISTATE BANK USING MACHINE LEARNING ALGORITHMS Hui Shan Hon1, Khai Wah Khaw2*, XinYing Chew3, Wai Peng Wong4 1,2* School of Management, Universiti Sains Malaysia, 11800 Minden, Pulau Pinang, Malaysia 3 School of Computer Sciences, Universiti Sains Malaysia, 11800 Minden, Pulau Pinang, Malaysia 4 School of Information Technology, Monash University, Malaysia Campus, Selangor, Malaysia [1](1honhuishan@gmail.com)[honhuishan@gmail.com](1honhuishan@gmail.com),2*[khaiwah@usm.my](khaiwah@usm.my), [3](3xinying@usm.my)[xinying@usm.my](3xinying@usm.my), [4](4waipeng.wong@monash.edu)[waipeng.wong@monash.edu](4waipeng.wong@monash.edu) ABSTRACT Customer churn is defined as the tendency of customers to cease doing business with a company in a given period. ABC Multistate Bankfaces the challenges to hold clients. The purpose of this study is to apply machine learning algorithms to develop the most effective model for predicting bank customer churn. In this study, six supervised machine learning methods, K-Nearest Neighbors, Support Vector Machine, Naïve Bayes, Decision Tree, Random Forest, and Extreme Gradient Boosting (XGBoost), are applied to the churn prediction model using Bank Customer Data of ABC Multistate Bank obtained from Kaggle. The results showed that XGBoost outperformed the other six classifiers, with an accuracy rate of 84. 76%, anF1 score of 56.95%, and a ROC curve graph of 71.64%. The bank may use XGBoost model to accurately identify customers who are at risk of leaving, concentrate their efforts on them, and possibly make a profit. Future research should focus on various machine learning approaches for determining the most accurate models for bank customer churn datasets. Keywords: Bank Customer Churn, Machine Learning, Supervised Machine Learning. Received for review: 19-06-2023; Accepted: 24-08-2023; Published: 10-10-2023 DOI: 10.24191/mjoc.v8i2.21393 1. Introduction Nowadays, due to the diversification of sales strategies brought about by technological advancements and the ensuing severe market competition, businesses are now required to move their attention from their products to their customers, as the customer is considered the real ruler of the market (Khodabandehlou & Rahman, 2017) . Customer churn (customer attrition), defined as the tendency of customers to cease doing business with a company in a given period (Chandar et al., 2006), has become a significant issue for ABC Multistate Bank. Churn in the banking sector is not only due to fierce competition but also due to increased choices available to customers, causing banks to feel pressure to maintain customer relationships in the long term. Predicting the customer churn rate in a business is important as acquiring new customers is more costly, sometimes five times the cost of retaining existing customers (Hung et al., 2006) . Predicting the likelihood of customer churn will serve as a guide for the bank in developing different marketing strategies for different client bases while satisfying their needs. Banks might develop proactive marketing campaigns to keep their existing customers from leaving. As a result of predicting customer churn, the bank can build a blueprint for its future revenue and assist its businesses in identifying and improving areas where customer service is lacking. Data mining is the process of examining large datasets to find hidden patterns, relationships, or even anomalies in the data (Pisal, 2022) . Machine learning is a subfield of artificial intelligence (AI) and computer science that uses data and algorithms to replicate how humans learn, gradually improving its accuracy. It can be used in a wide range of fields, including pavement performance prediction (Marcelino et al., 2021), spectroscopy (Meza Ramirez et al., 2021), medical field (Kaur & Kumari, 2022), machining industries (Cica et al.,2020), employee promotion prediction (Shafie et al., 2023) and in finance (M

FAQs

What is customer churn in this study’s context?

Customer churn is defined as the tendency of customers to cease doing business with a company within a given period.

Which machine learning algorithms are used to predict churn?

The study applies K-Nearest Neighbors, Support Vector Machine, Naïve Bayes, Decision Tree, Random Forest, and Extreme Gradient Boosting (XGBoost).

What result shows XGBoost’s advantage?

XGBoost outperforms the other classifiers, reporting about 84.76% accuracy, a 56.95% F1 score, and an ROC value around 71.64%.

Aug 4, 2026

Upload by Liam