Partial Dependency Profiles are averages from Ceteris Paribus Profiles. Function partial_dependency calls ceteris_paribus and then aggregate_profiles.

partial_dependency(x, ...)

# S3 method for explainer
partial_dependency(x, variables = NULL, N = 500,
  variable_splits = NULL, grid_points = 101, ...,
  variable_type = "numerical")

# S3 method for default
partial_dependency(x, data, predict_function = predict,
  label = class(x)[1], variables = NULL, grid_points = 101,
  variable_splits = NULL, N = 500, ..., variable_type = "numerical")

# S3 method for ceteris_paribus_explainer
partial_dependency(x, ...,
  variables = NULL)

Arguments

x

an explainer created with function DALEX::explain(), an object of the class ceteris_paribus_explainer or or a model to be explained.

...

other parameters

variables

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

N

number of observations used for calculation of partial dependency profiles. By default 500.

variable_splits

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

grid_points

number of points for profile. Will be passed to calculate_variable_split.

variable_type

a character. If numerical then only numerical variables will be calculated. If categorical then only categorical variables will be calculated.

data

validation dataset, will be extracted from x if it's an explainer NOTE: It is best when target variable is not present in the data

predict_function

predict function, will be extracted from x if it's an explainer

label

name of the model. By default it's extracted from the class attribute of the model

Value

an object of the class aggregated_profiles_explainer

Details

Find more detailes in the Partial Dependence Profiles Chapter.

References

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

Examples

library("DALEX") 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_glm <- partial_dependency(explain_titanic_glm, N = 50, variables = c("age", "fare")) head(pdp_glm)
#> Top profiles : #> _vname_ _label_ _x_ _yhat_ _ids_ #> 1 fare lm 0.0000000 0.3433783 0 #> 2 age lm 0.1666667 0.4136692 0 #> 3 age lm 2.0000000 0.4112289 0 #> 4 age lm 4.0000000 0.4085750 0 #> 5 fare lm 6.1793080 0.3512579 0 #> 6 age lm 7.0000000 0.4046102 0
plot(pdp_glm)
# \donttest{ library("randomForest") model_titanic_rf <- randomForest(survived ~., data = titanic_imputed)
#> Warning: The response has five or fewer unique values. Are you sure you want to do regression?
explain_titanic_rf <- explain(model_titanic_rf, data = titanic_imputed[,-8], y = titanic_imputed[,8], verbose = FALSE) pdp_rf <- partial_dependency(explain_titanic_rf, variable_type = "numerical") plot(pdp_rf)
pdp_rf <- partial_dependency(explain_titanic_rf, variable_type = "categorical") plotD3(pdp_rf, variable_type = "categorical", label_margin = 80, scale_plot = TRUE) # }