Churn Prediction in Telecom Using VNS Algorithm with Bootstrap Resampling Technique

Customer churn is often called customer attrition, or customer defect which is the rate at which customers are lost. Churn is important in the telecommunications industry because it directly influences the competitive position of the service provider. Customer churn prediction allows operators to have a period to remediate and implement a series of tactical retention measures before existing customers migrate to other operators. In this paper, we propose an algorithm based on variable neighborhood search using the mathematical programming technique to explore the customer churn prediction in telecom. We formulate the problem as a Linear Program. Due to the limited solver performance in large-scale instance processing, we propose an efficient heuristic based on variable neighborhood search concepts. The proposed algorithm finds a good solution in a reasonable time even for large instances. The effectiveness of this method is proven by experimental results obtained through the CrowdAnalytix public dataset.