| English | 中文 |
A linear predictor is a useful parsimonious comparison when additive relationships approximate the question. Review transformations and variable meanings, inspect residual structure and assess extrapolation. Correlated predictors can make coefficient interpretation unstable even when prediction is adequate. Predictive fitting alone supplies neither causal effects nor inferential uncertainty. Compare complexity against the intended spatial/time holdout, not training fit.
Parameter reasoning and cited evidence: model comparison.
Only the Python ordinary linear predictor is currently enabled; regularized variants and R linear regression are not exposed by the registered adapters. Configuration, code calls and current checks belong to the matching skill.
Use the shared evaluation design and submit a model plan before generating experiment code.