%0 Journal Article
%T A Neuro T-Norm Fuzzy Logic Based System
%A Alex Tserkovny
%J Journal of Software Engineering and Applications
%P 638-663
%@ 1945-3124
%D 2024
%I Scientific Research Publishing
%R 10.4236/jsea.2024.178035
%X In this study, we are first examining well-known approach to improve fuzzy reasoning model (<i>FRM</i>) by use of the genetic-based learning mechanism [1]. Later we propose our alternative way to build <i>FRM</i>, which has significant precision advantages and does not require any adjustment/learning. We put together neuro-fuzzy system (<i>NFS</i>) to connect the set of exemplar input feature vectors (<i>FV</i>) with associated output label (target), both represented by their membership functions (<i>MF</i>). Next unknown <i>FV</i> would be classified by getting upper value of current output <i>MF</i>. After that the fuzzy truths for all <i>MF</i> upper values are maximized and the label of the winner is considered as the class of the input <i>FV</i>. We use the knowledge in the exemplar-label pairs directly with no training. It sets up automatically and then classifies all input <i>FV</i> from the same population as the exemplar <i>FV</i>s. We show that our approach statistically is almost twice as accurate, as well-known genetic-based learning mechanism <i>FRM</i>.
%K Neuro-Fuzzy System
%K Neural Network
%K Fuzzy Logic
%K Modus Ponnens
%K Modus Tollens
%K Fuzzy Conditional Inference
%U http://www.scirp.org/journal/PaperInformation.aspx?PaperID=135573