Year: 2020
Type: article
Abstract: Churn Prophecy is the most salient sphere in the telecom industry for planning, decision making and to know the future events.Due to the tremendous amount of data, prediction has been always a tangled work.The prominent objective of churn Prophecy model is to strongly classify the customers into clusters and predict the prospective churners.Efficient feature selection is also required for predicting the churners.In this paper, an overview of recent machine learning works for customer churn prediction has provided.This paper provides XG boost classifier which less focused in the previous works.XG boost classifier is applied on publicly available telecom dataset and experiential results are compared with KNN and Random Forest Classifiers.XG boost classifier performs superior out of three classifiers mentioned.The evaluated metrics such as Precision, Recall, F1-score, Support are calculated.Finally, it is observed that among all the features, Fiber Optic customers with greater monthly chargers has higher cognition for churn using XG boost classifiers.
Authors: J.Beschi Raja, S. Chenthur Pandian
Language: English
Subfield: Marketing
Domain: Social Sciences
Open Access status: diamond
OpenAlex ID: W3021028588