GenAugData: Generate Augmented Data

In ellenxtan/ifedtree: Tree-based Federated Learning Approach for Personalized Effect Estimation and Prediction

Generate Augmented Data

Description

Generate augmented data from coordinating site and localized models. Calculate OOB prediction for coordinating site data on its model and common prediction on other models.

Usage

GenAugData(coord_id, coord_df, fit_lst, covars)

Arguments

coord_id	Index of coordinating site.
coord_df	Data of coordinating site ('data.table').
fit_lst	A list of localized models.
covars	A vector of covariate names used.

Value

Augmented data generated from the coordinating site and localized models.

Examples

K <- 4
df_lst <- list()
fit_lst <- list()
for (k in 1:K) {
    df <- GenSimData(k)
    df_lst[[k]] <- df
    covars <- grep("^X", names(df), value=TRUE)
    fit_lst[[k]] <- grf::causal_forest(X=as.matrix(df[, covars, with=FALSE]), Y=df$Y, W=df$Z)
}
coord_id <- 1
coord_df <- df_lst[[coord_id]]
GenAugData(coord_id, coord_df, fit_lst, covars)

ellenxtan/ifedtree documentation built on March 28, 2023, 9:09 a.m.