Predicting Customer Churn in Telecom: A Review of Advanced ML-Based Methods and Applications

Abstract

The telecoms industry is developing rapidly, which has increased competition. Retaining customers is now a highly
significant strategic objective. This article discusses the latest findings from studies that employ machine learning (ML) and deep
learning (DL) to forecast the number of telecom customers who would switch. Based on the study, telecom operators are able to
determine if customers are going to cancel their subscriptions based on their demographic, service usage, and behaviour information.
This sequence deals with decision tree models, logistic regression, RNNs, support vector machines, and artificial neural networks
(ANNs). The study also explores the application of churn prediction in real-life scenarios to improve customer retention, enable
targeted marketing, and enhance service quality. The largest issues, such as data imbalance, explainability, and scalability, and
directions for future research that encompass hybrid and explainable AI approaches. This study highlights the essential role of MLdriven
churn prediction in ensuring customer loyalty and maximizing the performance of telecom businesses.