Methodologies used for Customer Churn Detection in Customer Relationship Management

Abstract

Customer Relationship Management (CRM) is essential for many business organizations looking to maximize their customer interactions via Machine Learning and Deep Learning strategies came to be reshaping how businesses communicate with their customers. This paper look over the writing on the avail of ML and DL approaches to optimize CRM, as well as a description of the techniques used and how they are applied to every CRM dimension and component. Further, various functional effects of recent CRM method advances in the fields of ML and DL are examined. This article describes a method for applying ML strategies to assist such a corporation in dealing with client churn. We investigate different machine learning and Deep Learning algorithms utilizing real information from the XYZ Insurance customer churn dataset, which is headquartered in Indonesia. The classifiers used in this paper are Decision Tree using feature selection (DT using forward selection), Nave Bayes (NB), and Artificial Neural Network (ANN). DT with forward selection delivers the greatest results, with 91.3111 percent accuracy and 0.970, trailed by ANN and NB. It is advised that XYZ Insurance adopt the Decision Tree approach for such customer churn dataset and in overall. The paper provides useful information for further research moreover CRM resources that want to better their critical and automated functions.