Classification methods comparison for customer churn prediction in the telecommunication industry

The need for telecommunication services has increased dramatically in schools, offices, entertainment, and other areas. On the hand, competition between companies is getting tougher. Customer churn one of areas that each company gains more competitive advantage. This paper proposes a comparison several classification methods to make prediction whether customers cancel subscription service by highlighting key factors customer or not. non-trivial due urgent requirements from industry infer most appropriate techniques analyzing their churn. often huge commercial value. result shows Artificial Neural Network (ANN) can predict with an accuracy 79%, Support Vector Machine (SVM) 78% accuracy, Gaussian Naïve Bayes, K-Nearest Neighbor (KNN) 75% while Decision Tree 70% accuracy. Moreover, technique highest F-Measure Bayes 65% lowest 49%. Hence, ANN are two high recommendation industry.