%0 Conference Paper %1 IOEGC-12-171-12248 %A Shrestha, Sagar Maan %A Shakya, Aman %B Proceedings of 12th IOE Graduate Conference %D 2022 %K myown %P 1310 -- 1315 %T An XGBoost Based Ensemble Model for Customer Churn Prediction in Telecommunications Industry %U http://conference.ioe.edu.np/publications/ioegc12/IOEGC-12-171-12248.pdf %V 12 %X Telecommunications Industry are one of the fastest evolving business sectors. They demand huge financial investment even at the onset of business, unlike others. And the repay of their investment is driven by the number of customers they have garnished over time. With the increasing global and national competition this has become a major challenge for all the Telecommunication companies to retain their existing customers. Therefore, the telecommunication operators have major concern over identifying customers who are at risk of churning all the time. An analysis of call detail records inside the telecom will provide an insight on how customer behaviors are affected by the available services and provide an analysis part on whether they will be potential churners. In this research paper, we propose an ensemble-based Machine Learning model with the help of Stacking Technique on the Logistic Regression and Random Forest algorithms as base classifiers and XGBoost as final classifier for churn prediction using call detail records of a fictional telecommunication company that contains 7043 records of customers. The performance metrics like accuracy and F1-Score thus, obtained by this proposed model on this publicly available dataset are 80.88% and 62.69% respectively.

@inproceedings{IOEGC-12-171-12248, abstract = {Telecommunications Industry are one of the fastest evolving business sectors. They demand huge financial investment even at the onset of business, unlike others. And the repay of their investment is driven by the number of customers they have garnished over time. With the increasing global and national competition this has become a major challenge for all the Telecommunication companies to retain their existing customers. Therefore, the telecommunication operators have major concern over identifying customers who are at risk of churning all the time. An analysis of call detail records inside the telecom will provide an insight on how customer behaviors are affected by the available services and provide an analysis part on whether they will be potential churners. In this research paper, we propose an ensemble-based Machine Learning model with the help of Stacking Technique on the Logistic Regression and Random Forest algorithms as base classifiers and XGBoost as final classifier for churn prediction using call detail records of a fictional telecommunication company that contains 7043 records of customers. The performance metrics like accuracy and F1-Score thus, obtained by this proposed model on this publicly available dataset are 80.88% and 62.69% respectively.}, added-at = {2023-05-07T17:47:36.000+0200}, author = {Shrestha, Sagar Maan and Shakya, Aman}, biburl = {https://www.bibsonomy.org/bibtex/2432bd64a957b12385b9456c36cb6b501/amanshakya}, booktitle = {Proceedings of 12th IOE Graduate Conference}, interhash = {e75818e7874de921baab9dd7618e9b71}, intrahash = {432bd64a957b12385b9456c36cb6b501}, keywords = {myown}, month = {October}, organization = {Institute of Engineering, Tribhuvan University, Nepal}, pages = {1310 -- 1315}, timestamp = {2023-05-07T17:47:36.000+0200}, title = {An XGBoost Based Ensemble Model for Customer Churn Prediction in Telecommunications Industry}, url = {http://conference.ioe.edu.np/publications/ioegc12/IOEGC-12-171-12248.pdf}, volume = 12, year = 2022 }