Customer attrition is also known as customer churn, turnoveror defection which means the loss of clients or customers.In Telecommunications sector, they often use customer defection analysis and rates of their defection as a business metric topredict the clients earlier than their churn and try to retain them. This increases the business as cost ofretaining old customers is cost effective than cost involved in getting new customers.

Through this paper we proposed a modelfor predicting the customer churn earlier to win back defending clients using some of the algorithms likeLogistic Regression, Decision trees, Random Forest and some Boosting algorithms such as LightGBM, XGBoost.