Performance analysis of OKMSVM Learning Model with Existing Churn Prediction Models
Keywords:
Customer Churn Analysis, Telecommunication, Optimized Kernel Multiclass Support Vector Machine, Feature extraction using KPCA, Hybrid Two Level SVM modelAbstract
Due to a large clientele, the telecom industry generates enormous amounts of data every day. The telecom market is growing rapidly and is profitable from protocols, new computers, and communication skills in several countries. In this situation, data mining is essential to meet business needs, describe communication models, use sources efficiently, and improve facility excellence. Business analysts and decision-makers stressed that acquiring new clients is more expensive than keeping the ones you already have. They need to be aware of the causes of customer churn as well as any patterns in past churn customers' behaviour. In this article, we compared our proposed OKMSVM model with popular churn analysis models such as CNN, PBCCP and Ensemble learning model to have a better insight into the popular approaches for effective churn prediction.
