Prediction model development specification

Specify fitting tuning and complexity

Predictors serve prediction and need not represent causal effects. Tune and select models within the development procedure, then evaluate that procedure using data not used for the corresponding fitting decisions. Retain the full executable specification.

Flow diagram

Steps

Specify candidate model families

Choose regression or machine-learning approaches for the outcome and intended use.

Limit complexity and overfitting

Prespecify predictor handling, nonlinearities, regularization and interactions.

Separate tuning from evaluation

Define tuning objective and nested resampling or an equivalent valid procedure.

Fit the complete development pipeline

Record all data-dependent steps and reproducible software settings.

Freeze model specification

Retain coefficients or artifact version, preprocessing and prediction instructions.