How do you evaluate the performance of a machine learning model?

Model evaluation requires multiple metrics and validation strategies to ensure robust performance assessment across different scenarios.\n\n• Classification: Accuracy, precision, recall, F1-score, ROC-AUC, confusion matrix\n• Regression: MAE, MSE, RMSE, R-squared, MAPE\n• Cross-validation: K-fold, stratified, time-series splits\n• Business metrics: Revenue impact, user engagement, operational efficiency\n\nExample: Credit scoring model evaluation includes precision (minimize false positives), recall (catch actual defaults), AUC for ranking quality, and business metrics like profit per loan. Use stratified CV to maintain class balance and test on holdout data for final assessment.