Enhancing Performance of Relational Fuzzy Neural Networks with Square BK-Products
Warren Davis, Ladislav Kohout · Advances in intelligent systems research/Advances in Intelligent Systems Research · 2006
In this paper, we extend research done in max-min fuzzy neural networks in several important ways.We replace max and min operations used in the fuzzy operations by more general t-norms and co-norms, respectively.In addition, instead of the Łukasiewicz equivalence connective used in network of Reyes-Garcia and Bandler, we employ in our hybridization a variety of equivalence connectives.We explore the effectiveness of this network in the domain of phoneme recognition, and diabetes data.We find increased classification ability in many cases, as well as great potential for further expansion of the use of fuzzy operations in the field of pattern recognition.