Skip to content
Research Article Open access CC BY 4.0

Experimental Data Analysis to Recognize and Visualise the Factors Contributing to Bank Customer Churn Prediction Using Ensemble Learning Models

Subhani Shaik, S. Rishika, K. Nancy Reddy, M. Supraja

Asian Journal of Research in Computer Science · pp. 480–495 · Published 2 May 2025

10.9734/ajrcos/2025/v18i5668

Abstract

Customer churn happens when a client breaks down using a specific corporation's services and goods. It distresses the profit of industries heavily on their revenues. A novel customer acquisition is valuable up to five times more than absorbing an existing one. The target of this paper comprehend and predict customer churn in the banking sector. The bank wants to create more value out of its customer data. Analyse the data and propose how internal and external utilization of the analysis results increases the bank's revenues. We focus on exploratory data analysis to recognize and visualize features causative of client churn. This analysis predicts the purpose of constructing machine models. These models perform classification tasks for the given dataset. Multiple models were tested in our present research for bank customer churn prediction. Among all models, the LGBM model predicts the highest accuracy, 82.1%.

Exploratory data analysis identify visualize bank customer churn prediction ensemble machine learning models

Cited by 4

An Enhanced Multi-Task Deep Learning Framework for Joint Prediction of Customer Churn and Downsell

Qiang Zhang, Lihong Zhang, Yanfeng Chai · Applied Sciences · 2026

LightGBM for Customer Churn Prediction in Subscription-Based Services

Abdullah Elkhoudary, B. Almansour, Priya Vij · International Conference on Electrical Engineering and Informatics · 2025

Showing 2 of 4 known citations — external sources report more than can currently be individually listed.

Article metrics

Real usage data collected on this platform.

30

Page views

0

PDF downloads

4

Outbound clicks

4

Citations

Views over time

Views by country

Approximate, from request IP at view time — not citizenship or institution. Countries with fewer than 5 views are grouped as "Other".

Traffic sources

Referring site, by host.

Views and downloads exclude known bots/crawlers. Citations combines this platform's own DOI-resolved index with each external source's own reported total — see Cited by above for individually listed citing works. Last refreshed 0 seconds ago.