Application of Machine Learning in Predicting Customer Satisfaction of Telecom Service Providers

Abstract

Call drop in the cellular networking decides the top players of the industry where the most rated cellular network could sustain in the market and the least rated could face even bankruptcy. Customer satisfaction towards call quality is an important factor that any telecom organization must focus in order to survive successfully in the market in the long run. This research has demonstrated on how machine learning can be applied to predict customer satisfaction towards call drop quality of various telecom service providers. The dataset was acquired from Kaggle and the developed machine learning model has been trained and tested using the dataset. The research uses Random Forest Classifier for classification. The parameters under focus are the call rating, call drop, and the number of subscribers. The study focused on three months of data samples as the timeline (September to November) to predict and estimate customer satisfaction of cellular networking through developing a machine learning model that examined the call quality during a call, with the rating provided by the customers (call rating) at the call termination. The developed model was measured for its performance through recall, precision, accuracy and F1-score. The results obtained had rendered 91% accuracy in terms of predicting the customer satisfaction of call quality.