Attribute model predictions to input features using SHAP values.
npx skills add https://github.com/Clay-HHK/claude-config --skill shap-clay-hhk
Please help me install this Agent Skill. Skill: shap Source: https://github.com/Clay-HHK/claude-config/tree/main/skills/shap Command: npx skills add https://github.com/Clay-HHK/claude-config --skill shap-clay-hhk
SHAP provides a principled approach to attribute model predictions to individual input features, enabling transparent interpretation of complex models.
Explain a trained model's predictions by computing SHAP values and visualizing the results.
High-intent search queries and answers about installing and using this skill.
To explain machine-learning model predictions, SHAP attributes outputs to input features by applying specific explainers for tree-based, deep learning, and linear models to compute feature impact.
Feature attribution computes individual feature impacts on model predictions, enabling transparent model debugging and bias analysis by highlighting how specific inputs drive outputs across different model architectures.
Yes, SHAP explainers support both global and local interpretability for tree-based, deep learning, and linear models, allowing you to inspect overall feature impact and individual prediction attributions.
Yes, computing SHAP feature attributions requires handling background data correctly and specifying the proper model-output configuration to ensure accurate explanations across different model architectures.
The best way to visualize feature impact is by generating interactive and static visualizations from computed SHAP values, which inspect relationships and feature attributions for debugging and model comparison.
ManfronEnricoCompute SHAP values and visualizations for tree, deep, and linear models.
LogauaEngstromExplain model predictions by attributing feature contributions with SHAP values.
youyinnnExplain machine learning model predictions using SHAP values and plots.
tondevrelAnalyze scikit-learn model feature importance with SHAP, LIME, and partial dependence plots.
