Customer Churn Prediction in the Telecommunication Industry Using Multiple Machine Learning Algorithms

In the telecommunications sector, financial losses can be substantial due to customer churn. This study explores the prediction of churn in the industry using four popular machine learning algorithms: Logistic Regression, Random Forest, Gradient Boosting, and Histogram Gradient Boosting. A dataset comprising 7043 samples and 21 features, including demographic details (such as gender and partner information) and account-related information (like contract type, payment method, monthly charges, and customer service details), was utilized for model training and evaluation. The primary goal is to identify the most effective machine learning algorithm for forecasting customer attrition.The initial step involved data preparation, followed by the division of the dataset into testing and training sets. Subsequently, the four machine learning algorithms were employed to train and evaluate the predictive models. The results highlight the impact of data balancing on outcomes, demonstrating that all four algorithms effectively predict customer churn. Among them, Histogram Gradient Boosting outperformed the others with an accuracy of 85.56%, while Random Forest, Gradient Boosting, and Logistic Regression achieved accuracies of 85.41%, 84.73%, and 81.45%, respectively.The study concludes that machine learning algorithms are proficient in predicting telecom customer churn. Notably, Histogram Gradient Boosting emerged as the most effective algorithm for this task. Key predictors of customer churn include factors such as contract type, tenure, and monthly charges.