Remora Optimization with Machine Learning Driven Churn Prediction for Business Improvement

Abstract

Customer Relationship Management (CRM) is a complete approach to constructing, handling, and establishing loyal and long-lasting customer relationships. It is mostly acknowledged and widely executed for distinct domains, e.g., telecom, retail market, banking and insurance, and so on. A major objective is customer retention. The churn methods drive to recognize early churn signals and identify customers with an enhanced possibility to leave voluntarily. Machine learning (ML) techniques are presented for tackling the churning prediction difficult. This paper presents a Remora Optimization with Machine Learning Driven Churn Prediction for Business Improvement (ROML-CPBI) technique. The aim of the ROML-CPBI technique is to forecast the possibility of customer churns in the business sector. The working of the ROML-CPBI technique encompasses two major processes namely prediction and parameter tuning. At the initial stage, the ROML-CPBI technique utilizes multi-kernel extreme learning machine (MKELM) technique for churn prediction purposes. Secondly, the RO algorithm is applied for adjusting the parameters related to the MKELM model and thereby results in enhanced predictive outcomes. For validating the greater performance of the ROML-CPBI technique, an extensive range of experiments were performed. The experimental values signified the improved outcomes of the ROML-CPBI technique over other ones.