Function plotD3.feature_importance_explainer plots dropouts for variables used in the model. It uses output from feature_importance function that corresponds to permutation based measure of feature importance. Variables are sorted in the same order in all panels. The order depends on the average drop out loss. In different panels variable contributions may not look like sorted if variable importance is different in different models.

# S3 method for feature_importance_explainer
plotD3(
  x,
  ...,
  max_vars = NULL,
  show_boxplots = TRUE,
  bar_width = 12,
  split = "model",
  scale_height = FALSE,
  margin = 0.15,
  chart_title = "Feature importance"
)

Arguments

x

a feature importance explainer produced with the feature_importance() function

...

other explainers that shall be plotted together

max_vars

maximum number of variables that shall be presented for for each model. By default NULL which means all variables

show_boxplots

logical if TRUE (default) boxplot will be plotted to show permutation data.

bar_width

width of bars in px. By default 12px

split

either "model" or "feature" determines the plot layout

scale_height

a logical. If TRUE, the height of plot scales with window size. By default it's FALSE

margin

extend x axis domain range to adjust the plot. Usually value between 0.1 and 0.3, by default it's 0.15

chart_title

a character. Set custom title

Value

a r2d3 object.

References

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

Examples

library("DALEX") library("ingredients") lm_model <- lm(m2.price ~., data = apartments) explainer_lm <- explain(lm_model, data = apartments[,-1], y = apartments[,1], verbose = FALSE) fi_lm <- feature_importance(explainer_lm, loss_function = DALEX::loss_root_mean_square, B = 1) head(fi_lm)
#> variable mean_dropout_loss label #> 1 _full_model_ 279.3262 lm #> 2 construction.year 279.2987 lm #> 3 no.rooms 289.6181 lm #> 4 floor 503.1583 lm #> 5 surface 620.3488 lm #> 6 district 990.7506 lm
plotD3(fi_lm) # \donttest{ library("ranger") rf_model <- ranger(m2.price~., data = apartments) explainer_rf <- explain(rf_model, data = apartments[,-1], y = apartments[,1], label = "ranger forest", verbose = FALSE) fi_rf <- feature_importance(explainer_rf, loss_function = DALEX::loss_root_mean_square) head(fi_rf)
#> variable mean_dropout_loss label #> 1 _full_model_ 144.1809 ranger forest #> 2 no.rooms 300.9969 ranger forest #> 3 construction.year 378.9817 ranger forest #> 4 floor 426.2261 ranger forest #> 5 surface 521.2871 ranger forest #> 6 district 774.9314 ranger forest
plotD3(fi_lm, fi_rf) plotD3(fi_lm, fi_rf, split = "feature") plotD3(fi_lm, fi_rf, max_vars = 3, bar_width = 16, scale_height = TRUE) plotD3(fi_lm, fi_rf, max_vars = 3, bar_width = 16, split = "feature", scale_height = TRUE) plotD3(fi_lm, margin = 0.2) # }