Code and our presententation and collaboration during the Artificial Intelligence Seminar organised in Barcelona by the UPC & URV Universities. SHAP/Shapley Values
This repository provides an implementation and examples of SHAP/Shapley values, a popular method for interpreting the predictions of machine learning models. SHAP values are based on cooperative game theory and provide a unified framework for understanding the contribution of each feature to a model's output.
What are SHAP/Shapley Values? SHAP/Shapley values are a concept from cooperative game theory that aims to fairly allocate the contribution of each player in a cooperative game. In the context of machine learning models, SHAP values provide an explanation of how much each feature contributes to the model's prediction for a specific instance. They offer a comprehensive way to understand the importance and impact of individual features on model outputs.
This repository includes an implementation of SHAP/Shapley values for various types of machine learning models, including:
Linear models Tree-based models (e.g., decision trees, random forests, gradient boosting) Deep neural networks Ensemble models Features Implementation of SHAP/Shapley values for different types of machine learning models. Support for both classification and regression problems. Integration with popular Python libraries, including scikit-learn and TensorFlow. Example notebooks demonstrating the usage and interpretation of SHAP values. Efficient computation of SHAP values using sampling or approximation techniques. Visualization tools to create insightful plots and summaries of SHAP values.