Predict Method for hierNest Objects — predict

Provides predictions from a fitted hierarchical model (`hierNest`) using new data.

Usage

predict_hierNest(
  object,
  newx,
  hier_info,
  type = c("link", "response", "coefficients", "nonzero", "class"),
  ...
)

Arguments

object: A fitted hierNest model object.
newx: A numeric matrix of new predictor values for prediction.
hier_info: A numeric matrix with hierarchical grouping information. First column is MDC-level grouping; second column is DRG-level grouping.
type: Character string specifying the type of prediction required. Options include "link", "response", "coefficients", "nonzero", and "class".
...: Additional arguments passed to lower-level prediction methods.

Value

Predictions based on the specified `type`. Typically, returns:

Numeric vector or matrix of predicted values (for "link" or "response").
Model coefficients (for "coefficients").
Nonzero coefficient indices (for "nonzero").
Class labels for categorical outcomes (for "class").

Details

This function prepares a hierarchical design matrix based on `hier_info`, constructs the required Khatri-Rao product, and reorganizes it before generating predictions from the provided `object`.