Untitled

Abstract

The business world is becoming increasingly saturated in today's competitive environment. There is a great deal of competition in the telecommunications industry, especially due to various vibrant service providers. As a result, they have had difficulty retaining their existing customers. As attracting new customers is much more costly than retaining current ones, now is the time to ensure the telecom industry maintains value by retaining customers over acquiring new ones. Numerous machine learning and data mining methods have been proposed in the literature to predict customer churners using heterogeneous customer records over the past decade. This research gives a brief idea on the Customer Churn problem, and explores how various machine learning techniques can be used to predict customer churn via models such as XGBoost, GradientBoost, AdaBoost, ANN, Logistic Regression and Random Forest, and also compare the effectiveness of the models in term of accuracy.