Over the last few decades, there has been an increasing need for input-output mapping models. While AI techniques, especially deep learning techniques are revolutionizing the business and technology worlds, there is an increasing need to address the problem of interpretability and to improve model transparency, performance, and safety. In this chapter, we concentrate on fuzzy rule-based models, where the interpretability of the system is clear. We propose a three-step method based on Łukasiewicz logic for the description of the rules and the fuzzy memberships to construct concise and highly comprehensible fuzzy rules. We apply a genetic algorithm to evolve the structure of the rules and then a gradient-based optimization to fine-tune the fuzzy membership functions. Using the squashing functions introduced in Chap. 7, we can approximate the operators and the memberships in the same way and we can efficiently calculate the derivatives of the membership functions. We also present applications of the model using the UCI machine learning database.
CITATION STYLE
Dombi, J., & Csiszár, O. (2021). Learning Rules. In Studies in Fuzziness and Soft Computing (Vol. 408, pp. 135–145). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-72280-7_8
Mendeley helps you to discover research relevant for your work.