Evaluating ML Models

Use this skill for rigorously assessing model performance, comparing alternatives, diagnosing issues, and optimizing hyperparameters.

When to use this skill

• Model training complete — need systematic performance assessment
• Comparing multiple models/algorithms — statistical model comparison
• Diagnosing overfitting/underfitting — bias-variance analysis
• Hyperparameter tuning — finding optimal configurations
• Selecting appropriate metrics — matching metrics to business objectives
• Experiment tracking — reproducible experimentation
• Production readiness check — validation before deployment

When NOT to use this skill

• Feature engineering and preprocessing → use engineering-ml-features
• Exploratory data analysis → use analyzing-data
• Building interactive data apps → use @building-data-apps