Home / Current Issue / Paper 1710393
Machine Learning for Telecom Customer Retention and Growth
Subject area: Science,Engineering and Technology · Area of research: Machine Learning
DOI: https://doi.org/10.64388/IREV9I2-1710393-6842
Abstract
Customer churn is one of the most pressing problems for telecom operators, directly impacting revenue and long-term profitability. This paper presents a comprehensive machine learning framework for churn prediction using a publicly available telecom dataset. We describe data preprocessing, feature engineering, imbalance handling, model training, hyperparameter tuning, and evaluation. Algorithms evaluated include Logistic Regression, Decision Trees, Random Forest, Support Vector Machine (SVM), and XGBoost. Experiments use stratified 5-fold cross-validation and metric-based assessment (accuracy, precision, recall, F1-score, and AUC). Results show that ensemble techniques, particularly Random Forest and XGBoost, outperform simpler models, achieving the best balance between precision and recall on imbalanced data. We discuss practical deployment considerations for telecom providers, limitations of the current study, and directions for future work including online learning, explainability, and cost-sensitive retention strategies.
Keywords
Customer churn, Telecom, Machine learning, Random Forest, XGBoost, Imbalanced data, Retention
How to cite this paper
@article{1710393,
author = {Kethavath Hanmanthu, K. Balakrishna Maruthiram},
title = {Machine Learning for Telecom Customer Retention and Growth},
journal = {Iconic Research And Engineering Journals},
year = {2025},
volume = {9},
number = {3},
pages = {29-33},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1710393.pdf},
abstract = {Customer churn is one of the most pressing problems for telecom operators, directly impacting revenue and long-term profitability. This paper presents a comprehensive machine learning framework for churn prediction using a publicly available telecom dataset. We describe data preprocessing, feature engineering, imbalance handling, model training, hyperparameter tuning, and evaluation. Algorithms evaluated include Logistic Regression, Decision Trees, Random Forest, Support Vector Machine (SVM), and XGBoost. Experiments use stratified 5-fold cross-validation and metric-based assessment (accuracy, precision, recall, F1-score, and AUC). Results show that ensemble techniques, particularly Random Forest and XGBoost, outperform simpler models, achieving the best balance between precision and recall on imbalanced data. We discuss practical deployment considerations for telecom providers, limitations of the current study, and directions for future work including online learning, explainability, and cost-sensitive retention strategies.},
keywords = {Customer churn, Telecom, Machine learning, Random Forest, XGBoost, Imbalanced data, Retention},
month = {September},
doi = {https://doi.org/10.64388/IREV9I2-1710393-6842}
}