| Summary: | In this dissertation I present the implementation of a speech recognizer based on a fuzzy relational neural network model. In the model, the input acoustic features are represented by their respective fuzzy membership values to linguistic properties. The membership values are calculated with II functions, and trapezoidal functions. === The weights of the connections between input and output nodes are described in terms of their fuzzy relations. The output values, during the learning phase, are obtained by the use of the max-min composition, and are given in terms of fuzzy class membership values. The learning algorithm used is a modified version of the gradient-descent back-propagation algorithm. === The classification of unknown patterns is made using different approaches, one of which is the relational square product. The results and a comparison with several systems are presented as well. === Source: Dissertation Abstracts International, Volume: 55-04, Section: B, page: 1521. === Major Professor: Wyllis Bandler. === Thesis (Ph.D.)--The Florida State University, 1994.
|
|---|
