Attribute model predictions to input features using SHAP values.
npx skills add https://github.com/andikarachman/data-science-plugin --skill shap-andikarachman
Please help me install this Agent Skill. Skill: shap Source: https://github.com/andikarachman/data-science-plugin/tree/main/skills/shap Command: npx skills add https://github.com/andikarachman/data-science-plugin --skill shap-andikarachman
SHAP-based explanations translate opaque model predictions into human-understandable attributions, enabling trust, debugging, and governance of machine learning systems.
Train a model, create a SHAP explainer for your model type, and generate SHAP values and visualizations for a sample of your data.
High-intent search queries and answers about installing and using this skill.
SHAP values explain model predictions by attributing the output to individual input features. You create a SHAP explainer for your specific model type, compute the SHAP values, and generate visualizations like beeswarm or waterfall plots for global and local interpretability.
SHAP explainers support deep learning, tree-based, and linear models through automatic explainer selection. TreeExplainer, DeepExplainer, LinearExplainer, and KernelExplainer cover various model architectures to provide feature attribution and model interpretability.
Visualizing feature importance for individual predictions is best done using SHAP waterfall and scatter plots. These visualizations translate opaque model predictions into human-understandable attributions, allowing you to identify which features drive predictions for specific samples.
SHAP supports fairness analysis and production deployment workflows by providing guidance on caching explainers, selecting background data, and integrating explanations into dashboards. This enables model governance, trust, and debugging in production-ready pipelines.
Handling background data for SHAP values involves selecting appropriate reference datasets to compute feature attributions accurately. The Skill provides guidance on background data handling and automatic explainer selection to ensure reliable model interpretability across different tasks.
youyinnnExplain machine learning model predictions using SHAP values and plots.
Yezez9Compute SHAP feature attributions and generate explanation plots for machine learning models.
Weiwei-MaoCompute SHAP feature attributions and generate explanation plots for machine learning models.
sunLeeeDocument feature importance, SHAP, and partial dependence for machine learning models.
