A Comparison of Machine Learning Algorithms for Customer Churn Prediction

Today's fiercely competitive business environment has given significant importance to customer churn, a term used for the loss of customers, which possesses a significant challenge to organizations across various industries. To mitigate revenue loss and sustain growth, companies are increasingl...

Full description

Saved in:
Bibliographic Details
Published in:2023 6th International Conference on Advances in Science and Technology (ICAST) pp. 437 - 442
Main Authors: Pulkundwar, Parth, Rudani, Krishna, Rane, Omkar, Shah, Chintan, Virnodkar, Shyamal
Format: Conference Proceeding
Language:English
Published: IEEE 08.12.2023
Subjects:
Biological system modeling
churn
Churn Prediction
Classification algorithms
Customer
Data-driven Approaches
Gradient Boosting Algorithms
Machine Learning
Machine learning algorithms
Prediction algorithms
Predictive models
Random forests
Reviews
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Today's fiercely competitive business environment has given significant importance to customer churn, a term used for the loss of customers, which possesses a significant challenge to organizations across various industries. To mitigate revenue loss and sustain growth, companies are increasingly turning to machine learning (ML) algorithms for customer churn prediction. This review paper provides a concise examination of ML algorithms' role in predicting customer churn, a pivotal concern for businesses seeking to sustain growth and profitability. The review begins by underlining the significance of customer churn in today's competitive landscape, highlighting the impact of data-driven approaches in this context. The paper then explores various ML algorithms suitable for churn prediction and comparing the results to find out the most optimal algorithm for a few real-world scenarios, namely telecommunication, banking and ecommerce. The review found that Decision Tree Classification, Random Forest Classification, AdaBoost and XGBoost Classification algorithms were optimal for churn prediction. Additionally, the review covers the implementation of the findings in a churn prediction application.
DOI:10.1109/ICAST59062.2023.10455051