R/dalex_shortcut.R

In live: Local Interpretable (Model-Agnostic) Visual Explanations

#' Fit local model around the observation: shortcut for DALEX explainer objects
#'
#' @param explainer a model to be explained, preprocessed by the DALEX::explain function
#' @param observation a new observation for which predictions need to be explained
#' @param target_variable_name name of the response variablea as a character
#' @param n_new_obs Number of observation in the simulated dataset
#' @param local_model Character specyfing mlr learner to be used as a local model
#' @param select_variables If TRUE, variable selection will be performed while
#' fitting the local linear model
#' @param predict_type Argument passed to mlr::makeLearner() argument "predict.type"
#' while fitting the local model. Defaults to "response"
#' @param kernel_type Function which will be used to calculate distances from
#' simulated observation to explained instance
#' @param ... Arguments to be passed to sample_locally function
#' 
#' @return object of class live_explainer. More details in fit_explanation function help.
#'
#' @export
#' 
#' @examples
#' \dontrun{
#' data('wine')
#' library(randomForest)
#' library(DALEX)
#' rf <- randomForest(quality~., data = wine)
#' expl <- explain(rf, wine, wine$quality)
#' live_expl <- local_approximation(expl, wine[5, ], "quality", 500)
#' }
#'

local_approximation <- function(explainer, observation, target_variable_name,
                                n_new_obs,
                                local_model = "regr.lm",
                                select_variables = F,
                                predict_type = "response",
                                kernel_type = gaussian_kernel, ...) {
  
  neighbourhood <- sample_locally(explainer$data,
                                  observation,
                                  target_variable_name,
                                  n_new_obs,
                                  ...)
  with_predictions <- add_predictions(neighbourhood,
                                      explainer$model,
                                      predict_fun = explainer$predict_function)
  live::fit_explanation(with_predictions,
                        white_box = local_model,
                        selection = select_variables,
                        predict_type = predict_type,
                        kernel = kernel_type)
}