Currently, we are living in the era of big data, which is an innovative area of research and has been developing for a couple of decades. Due to the saturated situation in the market, telecommunication companies all over the world are trying to seek a competitive edge and to explore innovative and robust solutions for dealing with vulnerabilities of Big Data. It is a challenging task for them to dig deep into bulky data of customers, which may be in Tera Bytes or Pita Bytes, for extracting scientific insights to take decision for retaining its loyal customers in future. Therefore, it is essential for telecommunication companies to turn the bulky imbalanced dataset of customers into revenue. Because it is much expensive for the telecommunication companies to acquire more new customers rather than retaining the previous once. In this study, a novice robust and efficient fusion of data optimization techniques, based upon SMOTE, Ring-Theory and Particle Swarm Optimization, has hybridized with Heterogamous Ensemble Learning to deal with the data imbalance issue and to improve the overall performance of the customer churn prediction model. The proposed model has been named Optimized Customer Churn Prediction (OCCP) scored 0.940 Kappa Statistics, 0.0298 Mean Absolute Error, 0.173 Root Mean Square Error and 0.97 Accuracy. It is required to explore more innovative fusions of optimization techniques to improve accuracy a little bit more.