How is the accuracy of a predictive analytics model evaluated? What metrics such as accuracy, precision, recall, and F1-score are commonly used for model evaluation? How do evaluation techniques differ between classification and regression models? What role do validation methods like cross-validation play in measuring model performance? What challenges and limitations exist when interpreting model accuracy in real-world scenarios?