Customer Churn Prediction in Telecommunication System

Abstract

Churn Prediction is one of the world wide used analysis on Subscription Oriented Industries to analyze customer behaviors to predict the customers which are about to leave the service agreement from a company. We are using churn prediction model that uses classification to identify the churn customers. It is based on Machine Learning methods and algorithms and become so important for companies in today’s commercial conditions as gaining a new customer’s cost is more than retaining the existing ones. It is evaluated using metrics, such as accuracy, precision, recall, f-measure, and receiving operating characteristics (ROC) area. We are producing results using machine learning techniques like Random Forest, Decision Tree, Logistic Regression, XGBoost, Adaboost algorithms; out of all we got highest results in XGBoost. Churn Analysis on Telecommunication Industry in literature helps to present general information to readers about the frequently used data mining methods used, results and performance of the methods and shedding a light to further studies. To keep the review up to date, studies published in last five years and mainly last two years have been included.