liveBook · Manning

Interventional approach

As explained in Machine Learning for Tabular Data: XGBoost, Deep Learning, and AI

The interventional approach computes feature attributions by intervening on feature values according to an assumed causal dependence structure.

The interventional approach estimates SHAP feature contributions by breaking dependencies between features according to causal-inference assumptions. It is especially useful when tabular predictors are strongly collinear, though it can take longer and may no longer reconstruct the model’s original predictions exactly. section 7.3

Use the approach when:

  • Strong feature collinearity makes ordinary contribution estimates less reliable. section 7.3
  • More truthful-to-model contribution estimates matter more than exact prediction reconstruction. section 7.3
  • Longer SHAP computation times are acceptable for improved interpretive reliability. section 7.3
from shap import TreeExplainer

explainer = TreeExplainer(model_pipeline['xgb'], data=Xt, model_output='raw', feature_perturbation='interventional')
interventional_shap_values = explainer.shap_values(Xt)

Configures SHAP TreeExplainer with model output and interventional feature perturbation, then computes interventional SHAP values. section 7.3

How Machine Learning for Tabular Data: XGBoost, Deep Learning, and AI covers it

How deep it goes
Defines dependency-breaking perturbations and implementation settings, but stops short of deriving causal-inference rules.
Examples
Uses an XGBoost regression pipeline and a TreeExplainer configuration to calculate interventional SHAP values.
What it assumes
Assumes familiarity with SHAP values, tree models, feature perturbation, and causal-dependence assumptions.
What you can do after
Apply it via SHAP TreeExplainer to generate contribution estimates for correlated tabular-model features.
How it connects
Connects SHAP, TreeExplainer, causal inference, feature collinearity, and model prediction reconstruction.
Cover of Machine Learning for Tabular Data: XGBoost, Deep Learning, and AI
liveBookmanning.com
Read Machine Learning for Tabular Data: XGBoost, Deep Learning, and AI on liveBook Get the book at manning.com

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Machine Learning for Tabular Data: XGBoost, Deep Learning, and AI. The section links point to the parts of the book it draws on.