Predicting Telecom Customer Churn: An In-depth Evaluation of Machine Learning Algorithms

In such an advancing world, telecommunication companies confront the relentless problems of customer churn. Churning can be defined as cancelling the present company's subscription and buying a new service plan from another company. Churn Prediction is the technique that is used to detect these customers before they churn. For example, when JIO was launched, there was a significant change in the number of customers of Airtel, Idea, etc. Many of them churned and shifted to Jio as they offered economical plans. This paper focuses on utilizing machine learning algorithms like logistic regression, Extreme Gradient Boosting (XGBoost), Random Forest, etc to predict the customers who are likely to churn. We initiated the project by thoroughly exploring a Telecom Customer Churn dataset (WA_Fn-UseC_-Telco-Customer- Churn) available on Kaggle. This dataset includes several features like customer characteristics, service usage patterns and contract details. The insights from these features are important for the prediction of customer churn. Furthermore, we emphasize the development of a user-friendly front-end interface using StreamLit. This tool will enable telecom service providers to input customer information and receive real-time churn predictions, improving decision-making processes