compute_model_weights_from_preds: Function to compute (log) weights of component models from...

In reichlab/xgbstack: Flexible model stacking via gradient boosting for models that output predictive distributions.

Description

Function to compute (log) weights of component models from the "predictions" output by xgboost

Usage

compute_model_weights_from_preds(preds, log = FALSE)

Arguments

preds	predictions from xgboost in matrix form, with num_models columns and num_obs rows
log	boolean; return log of component model weights or just the weights

Value

a matrix with num_models columns and num_obs rows, with (log) weight for model m at observation i in entry [i, m]

reichlab/xgbstack documentation built on May 27, 2019, 4:54 a.m.