Predicting Customer Churn: A Comparison of Eight Machine Learning Techniques : A Case Study in an Indonesian Telecommunication Company

Abstract

Customer churn is considered as one of the most central problems for companies. Because of the direct influence on the revenues, particularly in high investment telecommunication companies. The companies are actively pursuing for developing methods to predict customers to churn. In addition, churn prediction modeling is exceptionally essential for companies to mitigate customers in order to reduce costs. This paper presents a comparative study of eight machine learning (ML) techniques to predict customer churn. They are ridge classifier, gradient booster, adaptive boosting (AdaBoost), bagging classifier, k-nearest neighbor (kNN), decision tree, logistic regression, and random forest. We applied those techniques to a dataset from a leading public telecommunication company in Indonesia. Finally, as a means of comparison, we tested a series of model evaluation metrics.