Generalizing Machine Learning Evaluation through the Integration of Shannon Entropy and Rough Set Theory

Olga Cherednichenko,Dmytro Chernyshov,Dmytro Sytnikov,Polina Sytnikova

2024-04-19

Abstract:This research paper delves into the innovative integration of Shannon entropy and rough set theory, presenting a novel approach to generalize the evaluation approach in machine learning. The conventional application of entropy, primarily focused on information uncertainty, is extended through its combination with rough set theory to offer a deeper insight into data's intrinsic structure and the interpretability of machine learning models. We introduce a comprehensive framework that synergizes the granularity of rough set theory with the uncertainty quantification of Shannon entropy, applied across a spectrum of machine learning algorithms. Our methodology is rigorously tested on various datasets, showcasing its capability to not only assess predictive performance but also to illuminate the underlying data complexity and model robustness. The results underscore the utility of this integrated approach in enhancing the evaluation landscape of machine learning, offering a multi-faceted perspective that balances accuracy with a profound understanding of data attributes and model dynamics. This paper contributes a groundbreaking perspective to machine learning evaluation, proposing a method that encapsulates a holistic view of model performance, thereby facilitating more informed decision-making in model selection and application.

Machine Learning

What problem does this paper attempt to address?

The paper aims to propose a new machine learning evaluation method by combining Shannon entropy and rough set theory, in order to address the problem of traditional evaluation metrics focusing only on prediction accuracy and ignoring the complexity of data feature interaction and learning process. This method provides a comprehensive framework by integrating the granularity analysis of rough sets and the quantification of information uncertainty of Shannon entropy, thereby revealing the inherent structure of data and the robustness of models while evaluating predictive performance. The experiments showed that this method can enhance the depth and breadth of model evaluation, contributing to wiser model selection and application decisions.