A Random Forest Churn Prediction Model: An Investigation of Machine Learning Techniques for Churn Prediction and Factor Identification in the Telecommunications Industry

Abstract

This study explores how machine learning techniques such as Random Forest (RF) algorithms can be used to predict and identify factors influencing churn in the telecommunications industry. This research uses a dataset from a local Italian telecommunications company to analyse customer behaviour and then implements the Random Forest algorithm to predict customer churn. Through methods such as feature engineering and parameter tuning, the results suggest that a relatively simple RF algorithm can provide a good prediction accuracy of churn churn and is able to identify the most important factors impacting churn. Further research is needed to analyse how these results can be applied to an enterprise setting in a more effective way.