Enhancing Customer Churn Prediction in Telecommunications: An Adaptive Ensemble Learning Approach