Predictions — reg_predict • mathmodels

Generates point predictions and confidence/ prediction intervals for new data.

Usage

reg_predict(model_result, n_new = 10, level = 0.95, newdata = NULL)

model_result: A fitted result from any reg_*() function.
n_new: Number of new observations to generate when newdata is not supplied. Predictor values are drawn from the empirical distributions of the input columns (normal approximation by mean/sd for numeric columns; levels sampled with observed proportions for factor columns).
level: Confidence/interval level. Defaults to 0.95.
newdata: Optional data frame of new predictor values. When supplied, predictions are made for exactly these rows and n_new is ignored.

A named list with:

predictions: Tibble with predicted values.
confidence_interval: Tibble with lower and upper bounds of the confidence interval.
prediction_interval: Tibble with lower and upper bounds of the prediction interval (linear models only).
newdata: The data frame actually used for prediction.