Customer churn prediction model: a case of the telecommunication market

Abstract The telecommunications market is well developed but characterized by oversaturation and high levels of competition. Based on this, the urgent problem to retain customers predict outflow customer base switching subscribers services competitors. Data Science technologies data mining methodology create significant opportunities for companies that implement analysis modeling development churn prediction models. research goals are compare different approaches methods construct models classify according probability their from company’s client potential who could stop use services. On example one leading Ukrainian telecommunication companies, article presents results classification models, such as C5.0, KNN, Neural Net, Ensemble, Random Tree, Net etc. All prepared in IBM SPSS Modeler have a level quality (the overall accuracy AUC ROC more than 90%). So, proves possibility feasibility using further loyalty company minimize consumer’s churn. key factors influencing identified form basis future optimization Implementation will help maintain loyalty, reduce increase business