Ensemble Classification using Balanced Data to Predict Customer Churn: A Case Study on the Telecom Industry

Published: 7 July 2022| Version 2 | DOI: 10.17632/ysdxb6557d.2
Contributors:
Omid Soleiman-garmabaki, Mohammad Hossein Rezvani

Description

This research examines the factors affecting customer churn in the telecom industry. In this regard, we use data mining classification methods, such as neural network, K-nearest neighbor, support vector machine, logistic regression, decision tree, and random forest. The results are analyzed using various criteria such as accuracy, precision, recall, F1-score, and ROC curve. After determining the successful classifiers, we combine them with the AdaBoost and XGBoost methods.

Files

Steps to reproduce

Please read the README file.

Institutions

  • Qazvin Islamic Azad University

Categories

Data Science

Licence