IBLM (version 1.0.2)

beta_corrections_derive: Compute Beta Corrections based on SHAP values

Description

Processes SHAP values in one-hot (wide) format to create beta coefficient corrections.

This includes:

• scaling shap values of continuous variables by the predictor value for that row

• migrating shap values to the bias for continuous variables where the predictor value was zero

• migrating shap values to the bias for categorical variables where the predictor value was reference level

Usage

beta_corrections_derive(
  shap_wide,
  wide_input_frame,
  iblm_model,
  migrate_reference_to_bias = TRUE
)

Value

A data frame with the booster model beta corrections in one-hot (wide) format

Arguments

shap_wide

Data frame containing SHAP values from XGBoost that have been converted to wide format by [shap_to_onehot()]

wide_input_frame

Wide format input data frame (one-hot encoded).

iblm_model

Object of class 'iblm'

migrate_reference_to_bias

Logical, migrate the beta corrections to the bias for reference levels? This applied to categorical vars only. It is recommended to leave this setting on TRUE

Examples

df_list <- freMTPLmini |> split_into_train_validate_test(seed = 9000)

iblm_model <- train_iblm_xgb(
  df_list,
  response_var = "ClaimRate",
  family = "poisson"
)

shap <- extract_booster_shap(iblm_model$booster_model, df_list$test)

wide_input_frame <- data_to_onehot(df_list$test, iblm_model)

shap_wide <- shap_to_onehot(shap, wide_input_frame, iblm_model)

beta_corrections <- beta_corrections_derive(shap_wide, wide_input_frame, iblm_model)

beta_corrections |> dplyr::glimpse()