Enhanced Customer Churn Prediction for CRM Using PBGL-WPLM and Optimized ANN Models

In today.s digital landscape, the information sector is the dominant service provider, results intense competition among them. To maintain strong market presence, service providers need to concentrate on customer churn, retention and customer satisfaction. Customer relationship management (CRM) strategies were particularly developed to improve , maintain and establish long term relationship with customer. Churn occurs when a customer switches from one service provider to another one. Predicting customer churn is challenging due to the lack of identifying churn reasons for customer churn. As a result, determining the efficiency of a prediction model is also dependent on how well the data can be interpreted in order to identify potential reasons for churn. Here Populace Based Gradual Learning — Wide and Profound Learning Models(PBGL-WPLM) method, Structure Optimized Simulated Annealing ANN, and Structure Optimized Hybrid SA-PBIL ANN methods are evaluated using the CRM dataset from the American media transmission associations and compared with novel technology called Populace Based Gradual Learning — Wide and Profound Learning Models Network Factorization(PBGL-WPLM-NF). Agitate is determined as the quantity of clients who leave the help somewhere in the range of 61 and 90 days subsequent to being tested.