Customer Churn Prediction of Telecom Company Using Machine Learning Algorithms

Abstract

We can’t escape the fact that using telecommunications has become a significant part of our everyday lives. Since the Covid-19 pandemic, the telecommunication industry has become crucial. Hence, the industry now enjoys growth opportunities. In this study, KNN, Random Forest (RF), AdaBoost, Logistic Regression (LR), XGBoost, and Support Vector Machine (SVM) are 6 supervised machine learning algorithms that will be used in this study to predict the customer churn of a telecom company in California. The goal of this study is to identify the classifier that predicts customer churn the most effectively. As evidenced by its accuracy of 79.67%, precision of 64.67%, recall of 51.87%, and F1-score of 57.57%, XGBoost is the overall most effective classifier in this study. Next, the purpose of this study is to identify the characteristics of customers who are most likely to leave the telecom company. These characteristics were discovered based on customers’ demographics and account information. Lastly, this study also provides the company with advice on how to retain customers. The study advises company to personalize the customer experience, implement a customer loyalty program, and apply AI in customer relationship management in retaining customers.