Global Variable Importance measure based on Partial Dependence profiles.

Source: R/global_variable_importance.R

This function calculate global importance measure.

global_variable_importance(profiles)

Arguments

profiles

data.frame generated by DALEX::model_profile() or DALEX::variable_profile()

Value

A data.frame of the class global_variable_importance. It's a data.frame with calculated global variable importance measure.

Examples



library("DALEX")
data(apartments)

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

explainer_rf <- explain(apartments_rf_model, data = apartmentsTest[,2:5],
                        y = apartmentsTest$m2.price)
#> Preparation of a new explainer is initiated
#>   -> model label       :  randomForest  (  default  )
#>   -> data              :  9000  rows  4  cols 
#>   -> target variable   :  9000  values 
#>   -> predict function  :  yhat.randomForest  will be used (  default  )
#>   -> predicted values  :  numerical, min =  2103.719 , mean =  3514.285 , max =  5364.833  
#>   -> model_info        :  package randomForest , ver. 4.6.14 , task regression (  default  ) 
#>   -> residual function :  difference between y and yhat (  default  )
#>   -> residuals         :  numerical, min =  -1252.104 , mean =  -2.761149 , max =  2139.647  
#>   A new explainer has been created!  

profiles <- model_profile(explainer_rf)

library("vivo")
global_variable_importance(profiles)
#>       variable_name  measure _label_model_
#> 1 construction.year 159.3080  randomForest
#> 2           surface 241.7430  randomForest
#> 3             floor 185.6088  randomForest
#> 4          no.rooms 185.7994  randomForest