An efficient churn prediction model using gradient boosting machine and metaheuristic optimization

Customer churn remains a critical challenge in telecommunications, necessitating effective churn prediction (CP) methodologies. This paper introduces the Enhanced Gradient Boosting Model (EGBM), which uses a Support Vector Machine with a Radial Basis Function kernel (SVM RBF ) as a base learner and...

Full description

Saved in:
Bibliographic Details
Published in:Scientific reports 2023-09, Vol.13 (1), p.14441-14441, Article 14441
Main Authors: AlShourbaji, Ibrahim, Helian, Na, Sun, Yi, Hussien, Abdelazim G., Abualigah, Laith, Elnaim, Bushra
Format: Article
Language:English
Subjects:
639/166/985
639/705/117
639/705/794
Comparative analysis
Humanities and Social Sciences
multidisciplinary
Optimization
Performance evaluation
Prediction models
Science
Science (multidisciplinary)
Statistical analysis
Telecommunications
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Customer churn remains a critical challenge in telecommunications, necessitating effective churn prediction (CP) methodologies. This paper introduces the Enhanced Gradient Boosting Model (EGBM), which uses a Support Vector Machine with a Radial Basis Function kernel (SVM RBF ) as a base learner and exponential loss function to enhance the learning process of the GBM. The novel base learner significantly improves the initial classification performance of the traditional GBM and achieves enhanced performance in CP-EGBM after multiple boosting stages by utilizing state-of-the-art decision tree learners. Further, a modified version of Particle Swarm Optimization (PSO) using the consumption operator of the Artificial Ecosystem Optimization (AEO) method to prevent premature convergence of the PSO in the local optima is developed to tune the hyper-parameters of the CP-EGBM effectively. Seven open-source CP datasets are used to evaluate the performance of the developed CP-EGBM model using several quantitative evaluation metrics. The results showed that the CP-EGBM is significantly better than GBM and SVM models. Results are statistically validated using the Friedman ranking test. The proposed CP-EGBM is also compared with recently reported models in the literature. Comparative analysis with state-of-the-art models showcases CP-EGBM's promising improvements, making it a robust and effective solution for churn prediction in the telecommunications industry.
ISSN:2045-2322
2045-2322
DOI:10.1038/s41598-023-41093-6