<ul><li>Decision-making process significantly influences predictions of machine learning models, especially in rule-based systems like Learning Fuzzy-Classifier Systems (LFCSs).</li><li>LFCSs combine evolutionary algorithms with supervised learning to optimize fuzzy classification rules, providing enhanced interpretability and robustness.</li><li>Introducing a novel class inference scheme for LFCSs based on Dempster-Shafer Theory of Evidence (DS theory) to handle uncertainty well and calculate belief masses for each class and the 'I don't know' state.</li><li>The proposed scheme demonstrates statistically significant improvements in test macro F1 scores across real-world datasets compared to conventional inference schemes, enhancing transparency, reliability, and generalizability of LFCSs.</li></ul>

A Class Inference Scheme With Dempster-Shafer Theory for Learning Fuzzy-Classifier Systems

Discover more