beezdemand (version 0.2.0)

predict.cp_model_lmer: Predict from a Mixed-Effects Cross-Price Demand Model

Description

Generates predictions from a mixed-effects cross-price demand model (of class cp_model_lmer). The function supports two modes:

Usage

# S3 method for cp_model_lmer
predict(object, newdata = NULL, pred_type = c("fixed", "random"), ...)

Value

A data frame containing all columns of newdata plus a column y_pred

with the corresponding predictions.

Arguments

object

A cp_model_lmer object (as returned by fit_cp_linear(type = "mixed", ...)).

newdata

A data frame containing at least an x column. For pred_type = "random", an id column is required. If absent, the function extracts unique ids from object$data and expands the grid accordingly. If no ids are available, a default id of 1 is used (with a warning).

pred_type

Character string specifying the type of prediction: either "fixed" (population-level) or "random" (subject-specific). The default is "fixed".

...

Additional arguments passed to the underlying predict function.

Details

"fixed"

Returns predictions based solely on the fixed-effects component (using re.form = NA).

"random"

Returns subject-specific predictions (fixed plus random effects) (using re.form = NULL).

Examples

# \donttest{
data(etm)
fit <- fit_cp_linear(etm, type = "mixed")
new_prices <- data.frame(x = c(2, 4, 8, 16, 32, 64))
predict(fit, newdata = new_prices, pred_type = "fixed")
# }