A study on the usefulness of the SHAP method in the process of hyperparameter tuning for machine learning methods
Defense Date:
This master’s thesis focuses on the analysis and assessment of the potential of the SHAP method (SHapley Additive exPlanations) in the context of tuning hyperparameters for machine learning methods. Originating from the field of XAI (eXplainable Artificial Intelligence), the SHAP method allows for a better interpretation of the results generated by machine learning models, highlighting the contribution of individual features to the decision-making process of the model. This work introduces an innovative approach to using the SHAP method in hyperparameter optimization. Instead of traditional methods like grid search or random search, a method leveraging SHAP was proposed, wherein the hyperparameter tuning process is based on information obtained from SHAP. This helps identify which hyperparameters have the most significant impact on the model’s performance in the context of a specific problem.