Churn prediction in telecommunication sector with machine learning methods


Şenyürek A., ALP S.

International Journal of Data Mining, Modelling and Management, vol.15, no.2, pp.184-202, 2023 (ESCI) identifier

  • Publication Type: Article / Abstract
  • Volume: 15 Issue: 2
  • Publication Date: 2023
  • Doi Number: 10.1504/ijdmmm.2023.131396
  • Journal Name: International Journal of Data Mining, Modelling and Management
  • Journal Indexes: Emerging Sources Citation Index (ESCI), Scopus, PASCAL, INSPEC, zbMATH
  • Page Numbers: pp.184-202
  • Keywords: churn analysis, CRM, customer relation management, machine learning, telecommunication
  • Yıldız Technical University Affiliated: Yes

Abstract

The aim of this study is to construct a model in which the subscribers are able to cancel their subscriptions in the telecommunication sector. In this context, it was aimed to select data, to prepare the preliminary preparation, to use machine learning method, performance criteria and measurement processes. According to logistic regression, artificial neural network, random forest and boosting method, potential churn subscribers were estimated. When the results of the study are examined, it is seen that the boosting method gives more accurate and successful results than the other methods. The most important factors causing customer churn was the period remaining until the end of the contract, tenure, which operator preferred the close relatives and the quality of the network.