ingredients (version 0.3.1)

aggregate_profiles: Aggregate Ceteris Paribus Profiles

Description

The function 'aggregate_profiles' calculates an aggregate of ceteris paribus profiles. It can be: Partial Dependency Profile (average across Ceteris Paribus Profiles), Conditional Dependency Profile (local weighted average across Ceteris Paribus Profiles) or Accumulated Local Dependency Profile (cummulated average local changes in Ceteris Paribus Profiles).

Usage

aggregate_profiles(x, ..., only_numerical = TRUE, groups = NULL,
  type = "partial", variables = NULL)

Arguments

x

a ceteris paribus explainer produced with function `ceteris_paribus()`

...

other explainers that shall be plotted together

only_numerical

a logical. If TRUE then only numerical variables will be plotted. If FALSE then only categorical variables will be plotted.

groups

a variable name that will be usef for grouping. By default 'NULL' which means that no groups shall be calculated

type

either 'partial'/'conditional'/'accumulated' for parital dependence, conditional profiles of accumulated local effects

variables

if not NULL then only `variables` will be presented

Value

an 'aggregated_profiles_explainer' layer

References

Predictive Models: Visual Exploration, Explanation and Debugging https://pbiecek.github.io/PM_VEE

Examples

# NOT RUN {
library("DALEX")
 
# }
# NOT RUN {
library("randomForest")
 model_titanic_rf <- randomForest(survived ~ gender + age + class + embarked +
                                    fare + sibsp + parch,  data = titanic)
 model_titanic_rf

 explain_titanic_rf <- explain(model_titanic_rf,
                           data = titanic[,-9],
                           y = titanic$survived)

selected_passangers <- select_sample(titanic, n = 100)
cp_rf <- ceteris_paribus(explain_titanic_rf, selected_passangers)
head(cp_rf)

pdp_rf_p <- aggregate_profiles(cp_rf, variables = "age", type = "partial")
pdp_rf_p$`_label_` <- "RF_partial"
pdp_rf_c <- aggregate_profiles(cp_rf, variables = "age", type = "conditional")
pdp_rf_c$`_label_` <- "RF_conditional"
pdp_rf_a <- aggregate_profiles(cp_rf, variables = "age", type = "accumulated")
pdp_rf_a$`_label_` <- "RF_accumulated"
plot(pdp_rf_p, pdp_rf_c, pdp_rf_a, color = "_label_")

pdp_rf <- aggregate_profiles(cp_rf, variables = "age",
                             groups = "gender")
head(pdp_rf)
plot(cp_rf, variables = "age") +
  show_observations(cp_rf, variables = "age") +
  show_rugs(cp_rf, variables = "age", color = "red") +
  show_aggreagated_profiles(pdp_rf, size = 3, color = "_label_")
# }