Churn Prediction using Ensemble Learning: An Analytical CRM Application

Abstract

Customer churn refers to the number of customers switching from one service provider to another service provider. Customer relationship management (CRM) is a business strategy that focuses on the needs of our customers by using advanced technologies. The goal is to improve business relationships and help companies to stay connected to customers. The main objective is to have an accurate churn prediction model that prevent churners, reduce churn rate, increase customer acquisition and increase retention of valuable customers. This will allow organizations to become proactive, anticipating outcomes and behaviors based upon the data. Combining the multiple classifiers in the existing system to create hybrid classifiers is a big challenge and designing the best ensemble to perform prediction is imprecise in this sector. The HYBRID FIREFLY algorithm exhibits higher accuracy of 86.34% by using random forest algorithm and gradient boosting method.