How to use breakDown package for models created with xgboost

In breakDown: Model Agnostic Explainers for Individual Predictions

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

This example demonstrates how to use the breakDown package for models created with the xgboost package.

library("breakDown")
if (requireNamespace("xgboost", quietly = TRUE)) {
  library(xgboost)

  model_martix_train <- model.matrix(left ~ . - 1, HR_data)
  data_train <- xgb.DMatrix(model_martix_train, label = as.numeric(HR_data$left))
  param <- list(objective = "reg:linear")

  HR_xgb_model <- xgb.train(param, data_train, nrounds = 50)
  HR_xgb_model 
}

Now we are ready to call the broken() function.

if (requireNamespace("xgboost", quietly = TRUE)) {
  library("breakDown")
  nobs <- model_martix_train[1L, , drop = FALSE]

  explain_2 <- broken(HR_xgb_model, new_observation = nobs, 
                      data = model_martix_train)
  explain_2
}

And plot it.

if (requireNamespace("xgboost", quietly = TRUE)) {
  library(ggplot2)
  plot(explain_2) + ggtitle("breakDown plot for xgboost model")
}

breakDown documentation built on Jan. 20, 2021, 5:06 p.m.