Customer Size Prediction Using Machine Learning Approach for Mobile Package

Abstract

Nowadays the telecom market is competitive and telecom operators launch various new service packages to meet customer needs and attract more customers as well. Ethio telecom is the only telecommunications service provider in Ethiopia. In the case of ethio telecom, as there is no an automated method for package preview, Machine Learning (ML) approach has been studied to predict customer size for new mobile packages. Three ML algorithms that are, ElasticNet regression, Extreme Gradient Boosting and Random Forest regression (RF) have been used to train the prediction models. To train the model, mobile package dataset has been constructed by integrating data from three different sources in ethio telecom. The sources are business support systems, marketing product catalog and mobile package post launch analysis results. As the study has showed, the RF model has outperformed the ElasticNet regression and Extreme Gradient Boosting models.