plot_interaction: SHAP Interaction value plot

In ModelOriented/treeshap: Compute SHAP Values for Your Tree-Based Models Using the 'TreeSHAP' Algorithm

SHAP Interaction value plot

Description

This function plots SHAP Interaction value for two variables depending on the value of the first variable. Value of the second variable is marked with the color.

Usage

plot_interaction(
  treeshap,
  var1,
  var2,
  title = "SHAP Interaction Value Plot",
  subtitle = ""
)

Arguments

treeshap	A treeshap object produced with treeshap(interactions = TRUE) function. treeshap.object.
var1	name or index of the first variable - plotted on x axis.
var2	name or index of the second variable - marked with color.
title	the plot's title, by default 'SHAP Interaction Value Plot'.
subtitle	the plot's subtitle. By default no subtitle.

Value

a ggplot2 object

See Also

treeshap for calculation of SHAP Interaction values

plot_contribution, plot_feature_importance, plot_feature_dependence

Examples

data <- fifa20$data[colnames(fifa20$data) != 'work_rate']
target <- fifa20$target
param2 <- list(objective = "reg:squarederror", max_depth = 5)
xgb_model2 <- xgboost::xgboost(as.matrix(data), params = param2, label = target, nrounds = 10)
unified_model2 <- xgboost.unify(xgb_model2, data)
inters <- treeshap(unified_model2, as.matrix(data[1:50, ]), interactions = TRUE)
plot_interaction(inters, "dribbling", "defending")

ModelOriented/treeshap documentation built on Jan. 27, 2024, 10:57 p.m.