Report Number: CS-TR-95-1550
Institution: Stanford University, Department of Computer Science
Title: Theory and Design of a Hybrid Pattern Recognition System
Author: Drakopoulos, John A.
Date: May 1995
Abstract: Pattern recognition methods can be divided into four different categories: statistical or probabilistic, structural, possibilistic or fuzzy, and neural methods. A formal analysis shows that there is a computational complexity versus representational power trade-off between probabilistic and possibilistic or fuzzy set measures, in general. Furthermore, sigmoidal theory shows that fuzzy set membership can be represented effectively by sigmoidal functions. Those results and the formalization of sigmoidal functions and subsequently multi-sigmoidal functions and neural networks led to the development of a hybrid pattern recognition system called tFPR. tFPR is a hybrid fuzzy, neural, and structural pattern recognition system that uses fuzzy sets to represent multi-variate pattern classes that can be either static or dynamic depending on time or some other parameter space. The membership functions of the fuzzy sets that represent pattern classes are modeled in three different ways. Simple sigmoidal configurations are used for simple patterns, a structural pattern recognition method is used for dynamic patterns, and multi-sigmoidal neural networks are used for pattern classes for which is difficult to obtain a formal definition. Although efficiency is a very important consideration in tFPR, the main issues are knowledge acquisition and knowledge representation (in terms of pattern class descriptions).