Advanced Customer Lifetime Value Predictive Model

Design a comprehensive Python script that predicts customer lifetime value (CLV) using machine learning, incorporating advanced feature engineering techniques. The model should handle multi-dimensional data from e-commerce transactions, including purchase frequency, monetary value, and recency. Implement both random forest and gradient boosting models, with cross-validation, and generate a performance comparison dashboard that shows predictive accuracy, feature importance, and potential revenue uplift. Include robust error handling for missing data and outlier detection.