Explain predictive models

You have a forest fit, but what does it say about a clinical prediction? Start with whether it converged, then choose the slice that answers what it predicts, which variables matter, and where a predictor changes the response.

Choose the next recipe

Your question	Recipe	Why this one
Did the forest settle as more trees were grown?	Random forest error convergence	Reads the out-of-bag error curve before interpreting the fit.
What does the forest predict for a patient or group?	Predicted response and survival	Shows fitted responses and survival estimates.
Which predictors contribute most to the fit?	Variable importance	Ranks variables by their contribution to prediction.
How did a variable contribute to an individual prediction?	SHAP attribution	Shows per-observation, per-variable attributions.
How does the predicted response change over a predictor?	Variable and partial dependence	Draws marginal and partial effects.
How well does a classification forest discriminate and calibrate?	ROC and Brier performance	Gives ROC and Brier views of classification performance.
Which variables survive rule-based selection?	Variable priority with varPro	Shows the variable-priority workflow.
What does a selected variable do across its range?	varPro partial dependence	Reads the partial effect after selection.
How do I inspect a forest with time-varying covariates?	Random Hazard Forests	Uses start-stop data when predictors and risk change with time.