ceterisParibus (version 0.4.2)

local_conditional_expectations: Local Conditional Expectation Explainer

Description

This explainer works for individual observations. For each observation it calculates Local Conditional Expectation (LCE) profiles for selected variables.

Usage

local_conditional_expectations(
  explainer,
  observations,
  y = NULL,
  variable_splits = NULL,
  variables = NULL,
  grid_points = 101
)

Value

An object of the class 'ceteris_paribus_explainer'. A data frame with calculated LCE profiles.

Arguments

explainer

a model to be explained, preprocessed by function `DALEX::explain()`.

observations

set of observarvation for which profiles are to be calculated

y

true labels for `observations`. If specified then will be added to local conditional expectations plots.

variable_splits

named list of splits for variables, in most cases created with `calculate_variable_splits()`. If NULL then it will be calculated based on validation data avaliable in the `explainer`.

variables

names of variables for which profiles shall be calculated. Will be passed to `calculate_variable_splits()`. If NULL then all variables from the validation data will be used.

grid_points

number of points for profile. Will be passed to `calculate_variable_splits()`.

Examples

library("DALEX")
 if (FALSE) {
library("randomForest")
set.seed(59)

apartments_rf_model <- randomForest(m2.price ~ construction.year + surface + floor +
      no.rooms + district, data = apartments)

explainer_rf <- explain(apartments_rf_model,
      data = apartments[,2:6], y = apartments$m2.price)

new_apartment <- apartments[1, ]

cp_rf <- ceteris_paribus(explainer_rf, new_apartment)
lce_rf <- local_conditional_expectations(explainer_rf, new_apartment)
lce_rf

lce_rf <- local_conditional_expectations(explainer_rf, new_apartment, y = new_apartment$m2.price)
lce_rf

# Plot LCE
sel_vars <- c("surface", "no.rooms")
plot(lce_rf, selected_variables = sel_vars)

# Compare ceteris paribus profiles with LCE profiles 
plot(cp_rf, selected_variables = sel_vars) + 
   ceteris_paribus_layer(lce_rf, selected_variables = sel_vars, color = "red")

}