Customer Churn Prediction In Telecommunication Industry Using Machine Learning Classifiers

Abstract

Customer churn is one of the main problems in telecommunication industry. This study aims to identify the factors that influence customer churn and develop an effective churn prediction model as well as provide best analysis of data visualization results. The dataset has been collected from Kaggle open data website. The proposed methodology for analysis of churn prediction covers several phases: data pre-processing, analysis, implementing machine learning algorithms, evaluation of the classifiers and choose the best one for prediction. Data preprocessing process involved three major action, which are data cleaning, data transformation and feature selection. Machine learning classifiers was chosen are Logistic Regression, Artificial Neural Network and Random Forest. Then, classifiers were evaluated by using performance measurement which are accuracy, precision, recall and error rate in order to find the best classifier. Based on this study, the output shows that logistic regression outperform compared to artificial neural network and random forest.