Churn Prediction Model for Pakistan Telecom Company using Machine Learning

Abstract

Over the last two decades, we have seen mobile telecommunication become the dominant communication medium. In many countries, especially developed ones, the market has reached a degree of saturation where each new customer must be won over from the competitors. Since the cost of winning a new customer is far greater than the cost of preserving an existing one, mobile carriers have been shifting considerable attention from customer acquisition to customer retention. As a result, churn prediction has emerged as a crucial mobile Business Intelligence (BI) application that aims at identifying customers who are about to transfer their business to a competitor (i.e., to churn). This aim of this project was to develop a churn prediction model using commonly used Machine Learning Algorithms. In order to analyze the trends and generate results, anonymous data pertinent to Telecom Industry was collected from a local telecom company. This data was later cleansed to remove all anomalies and missing entries. Well known classifier algorithms were used especially C5.0, Classification & Regression Tree and Discriminant Analysis. A separate model for each classifier was generated, trained and tested on the dataset in order to find patterns which helped in pointing out and predicting possible churners.