An adaptive analytics framework for customer retention through integrative feature optimization and ensemble learning

An adaptive analytics workflow is presented for customer churn prediction, combining Principal Component Analysis for dimensionality reduction, a hybrid Modified Particle Swarm Gravitational Search Optimization (MPSO-GSO) for feature selection and hyperparameter tuning, and an ensemble learning stage combining XGBoost and LightGBM through weighted voting. Applied to an e-commerce dataset, the complete framework achieves AUC = 0.99 and accuracy = 0.98, outperforming standalone XGBoost (AUC = 0.98) and LightGBM (AUC = 0.97). Stratified 5-fold cross-validation and paired t-tests confirm the statistical significance of this improvement (p < 0.01). Subsequent SHAP analysis interprets the feature contributions, demonstrating that this integrative, optimization-based approach substantially improves the quality of churn prediction. • Combine dimensionality reduction and feature optimization to enhance predictive modeling in customer retention. • Apply a hybrid ensemble of decision trees to improve churn prediction accuracy in business analytics. • Validate that an integrated modeling pipeline boosts decision-making performance across business variables. • Use interpretability tools to explain how customer features influence predictive model outcomes. • Demonstrate improved forecasting accuracy by merging optimization techniques with ensemble learning methods.