plot.aggregated_profiles_explainer: Plots Aggregated Profiles

In ingredients: Effects and Importances of Model Ingredients

Plots Aggregated Profiles

Description

Function plot.aggregated_profiles_explainer plots partial dependence plot or accumulated effect plot. It works in a similar way to plot.ceteris_paribus, but instead of individual profiles show average profiles for each variable listed in the variables vector.

Usage

## S3 method for class 'aggregated_profiles_explainer'
plot(
  x,
  ...,
  size = 1,
  alpha = 1,
  color = "_label_",
  facet_ncol = NULL,
  facet_scales = "free_x",
  variables = NULL,
  title = NULL,
  subtitle = NULL
)

Arguments

x	a ceteris paribus explainer produced with function aggregate_profiles()
...	other explainers that shall be plotted together
size	a numeric. Size of lines to be plotted
alpha	a numeric between 0 and 1. Opacity of lines
color	a character. Either name of a color, or hex code for a color, or _label_ if models shall be colored, or _ids_ if instances shall be colored
facet_ncol	number of columns for the facet_wrap
facet_scales	a character value for the facet_wrap. Default is "free_x".
variables	if not NULL then only variables will be presented
title	a character. Partial and accumulated dependence explainers have deafult value.
subtitle	a character. If NULL value will be dependent on model usage.

Value

a ggplot2 object

References

Explanatory Model Analysis. Explore, Explain, and Examine Predictive Models. https://ema.drwhy.ai/

Examples

library("DALEX")
library("ingredients")

model_titanic_glm <- glm(survived ~ gender + age + fare,
                         data = titanic_imputed, family = "binomial")

explain_titanic_glm <- explain(model_titanic_glm,
                               data = titanic_imputed[,-8],
                               y = titanic_imputed[,8],
                               verbose = FALSE)

pdp_rf_p <- partial_dependence(explain_titanic_glm, N = 50)
pdp_rf_p$`_label_` <- "RF_partial"
pdp_rf_l <- conditional_dependence(explain_titanic_glm, N = 50)
pdp_rf_l$`_label_` <- "RF_local"
pdp_rf_a<- accumulated_dependence(explain_titanic_glm, N = 50)
pdp_rf_a$`_label_` <- "RF_accumulated"
head(pdp_rf_p)
plot(pdp_rf_p, pdp_rf_l, pdp_rf_a, color = "_label_")


library("ranger")

model_titanic_rf <- ranger(survived ~., data = titanic_imputed, probability = TRUE)

explain_titanic_rf <- explain(model_titanic_rf,
                              data = titanic_imputed[,-8],
                              y = titanic_imputed[,8],
                              label = "ranger forest",
                              verbose = FALSE)

selected_passangers <- select_sample(titanic_imputed, n = 100)
cp_rf <- ceteris_paribus(explain_titanic_rf, selected_passangers)
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"

head(pdp_rf_p)
plot(pdp_rf_p)
plot(pdp_rf_p, pdp_rf_c, pdp_rf_a)

plot(cp_rf, variables = "age") +
  show_observations(cp_rf, variables = "age") +
  show_rugs(cp_rf, variables = "age", color = "red") +
  show_aggregated_profiles(pdp_rf_p, size = 2)

ingredients documentation built on Jan. 15, 2023, 5:09 p.m.