Logical operation based fuzzy MLP for classification and rule generation

Abstract

A fuzzy layered neural network for classification and rule generation is proposed using logical neurons. It can handle uncertainty and/or impreciseness in the input as well as the output. Logical operators, namely, t-norm T and t-conorm S involving And and Or neurons, are employed in place of the weighted sum and sigmoid functions. Various fuzzy implication operators are introduced to incorporate different amounts of mutual interaction during the back propagation of erros. In case of partial inputs the model is capable of querying the user for the more important feature information, if and when required. Justification for an inferred decision may be produced in rule form. The built-in And-Or structure of the network enables the generation of appropriate rules expressed as the disjunction of conjunctive clauses. The effectiveness of the model is tested on a speech recognition problem and on some artificially generated pattern sets.