Machine Learning Approaches to Predict New Mobile Internet Customers

Abstract

Globalization and liberalization of the economy dramatically shifted the nature of business competition. The emergence of new technology in business operations has intensified rivalry and generated new opportunities for service providers. In order to deal with increasing situations, businesses are turning their focus to maintaining current clients rather than acquiring new ones. This is more cost-effective and therefore needs fewer energy. In this article, future mobile internet customers are investigated on the basis of machine learning and deep learning strategies applied to consumer activity and usage knowledge, which can assist new mobile internet customers. This paper utilized consumer usage and similar knowledge from a telephone service provider to examine mobile internet customers in the telecommunications industry. XGBoost and Random Forest the decision tree ensembles are used as basic statistical machine learning models for the development of a binary mobile internet classifier. The implementation component was developed using Python, a state-of-the-art structured data processing platform for machine learning and data mining. Many ML and deep learning approaches such as (K-Nearest Neighbors KNN, logistic regression, Support Vector Machine SVM, and Deep Neural Network DNN) have been tested to achieve greater and more successful outcomes and results.