Article

Diagnosing Causal Credibility of Machine Learning Explanations: A Dual- SHAP Attribution Framework Applied to Social Survey Data

2026-08-20

Abstract excerpt

<title>Abstract</title> <p>Machine learning models increasingly inform social science research, yet SHAP feature attributions conflate correlational and causal relationships, potentially misleading interpretations. We propose a dual-SHAP framework comparing conditional and interventional SHAP values to diagnose causal credibility of feature attributions. The framework integrates model comparison, SHAP-based featu...

Topics

Open a Topic to create a Post that cites this publication.

Traditional Chinese Medicine Studies

Identifiers and source

Literature Corpus work: d64f2add-bb39-57bf-93f0-b0fbe89b379b
DOI: 10.21203/rs.3.rs-10687172/v1

Open publication

Related research

Semantic proximity does not establish scientific evidence.

Click a neighbor to travelStep 1 · 12 closest

Diagnosing Causal Credibility of Machine Learning Explanations: A Dual- SHAP Attribution Framework Applied to Social Survey DataDOI 10.21203/rs.3.rs-10687172/v1Read

Gathering, Evaluating, and Aggregating Social Scientific ModelsSemantic proximity Read

Comparative Evaluation of SHAP and LIME for Clinical Interpretability in Postoperative Cardiac Surgery Mortality Prediction ModelsSemantic proximity Read

DeepDiff-SHAP: Interpretable deep learning for subgroup-specific causal inference using conditional SHAPSemantic proximity Read

Quantifying Trust: Stability of SHAP in Pan-Cancer ModelsSemantic proximity Read

Causal AI for Data Scientists: A Framework to Automate Discovery of Causal Relationships in Noisy DatasetsSemantic proximity Read

Interpretable machine learning with tree-based shapley additive explanations: application to metabolomics datasets for binary classificationSemantic proximity Read

A Basic Trustworthy Machine Learning Framework for Early Diabetes DetectionSemantic proximity Read

IHCP: interpretable hepatitis C prediction system based on black-box machine learning modelsSemantic proximity Read

A Framework for Generating Counterfactual Explanations to Explain Black-Box ModelsSemantic proximity Read

Building Compressed Causal Models of the WorldSemantic proximity Read

Refocusing Algorithmic Fairness on Feature-Level Bias: A Diagnostic Approach Using Dutch EHR DataSemantic proximity Read

Multi-Domain Counterfactual Causal Graphs for Spurious Pathway Detection and Functional Risk EstimationSemantic proximity Read