Attribute input feature contributions to model predictions using SHAP values.
npx skills add https://github.com/mkurman/tamux --skill shap-mkurman
Please help me install this Agent Skill. Skill: shap Source: https://github.com/mkurman/tamux/tree/main/skills/scientific-skills/shap Command: npx skills add https://github.com/mkurman/tamux --skill shap-mkurman
SHAP-based explanations help you understand why a model made a specific prediction by attributing contributions to individual features using Shapley values.
Explain a single prediction by computing SHAP values and visualizing the contributions for the given input.
High-intent search queries and answers about installing and using this skill.
Model predictions are explained using SHAP values by attributing the contribution of each input feature to the final output, enabling both global feature importance analysis and local explanations for specific instances.
SHAP supports interpreting tree-based ensembles, deep networks, linear models, and black-box predictors using dedicated explainers like TreeExplainer, DeepExplainer, LinearExplainer, and KernelExplainer.
Feature attribution is visualized through SHAP workflows that generate plots like waterfall, beeswarm, bar, scatter, force, and heatmap charts to illustrate how individual features drive model decisions.
SHAP allows analyzing feature interactions and debugging model behavior by calculating interaction values, auditing fairness, and comparing model performance across different datasets.
Background data is required for baseline calculation when using KernelExplainer to interpret black-box predictors, ensuring accurate Shapley value approximations for your feature attributions.
ToqsickSearch, download, and upload models and datasets on Hugging Face Hub.
tomlupoExplain model predictions by attributing outputs to input features with SHAP values.
xiechyCompute SHAP values to attribute predictions to features across model types.
MassimilianoPiliWrite unit and integration tests for Angular components with Vitest or Jasmine.
