Title :
An assessment of conjunctoid theory and its potential utility
Author :
Ma, Keping ; Jannarone, Robert J. ; Gorman, John W.
Author_Institution :
Dept. of Electr. & Comput. Eng., South Carolina Univ., Columbia, SC, USA
Abstract :
Conjunctoids and their advantages to nonstatisticians are described by using clear examples and avoiding formal statistical concepts. Examples are given of conjunctoids that learn to recognize patterns, evaluate logical functions, filter out noise, and correct errors. Conjunctoid neurocomputing models can be implemented either as parallel architectures or as sequential architectures, both of which include learning and performance phases. Parameter estimation plays a key role during the learning phase. All parameters are stored in the parameter memory and are updated either in parallel or sequentially according to the architecture. Examples include 7-20-segment LED pattern recognition learning algorithms simulated on a SUN-4 and implemented on an NCUBE massively parallel processing computer. The results show that the conjunctoid neurocomputing algorithms can learn to recognize any binary pattern
Keywords :
computerised pattern recognition; learning systems; parallel algorithms; parameter estimation; statistics; 7-20-segment LED pattern recognition learning algorithms; NCUBE massively parallel processing computer; SUN-4; binary pattern recognition; conjunctoid neurocomputing models; conjunctoid theory; learning; logical functions; parallel architectures; sequential architectures; Computational modeling; Computer architecture; Computer simulation; Error correction; Filters; Light emitting diodes; Parallel architectures; Parallel processing; Parameter estimation; Pattern recognition;
Conference_Titel :
Southeastcon '90. Proceedings., IEEE
Conference_Location :
New Orleans, LA
DOI :
10.1109/SECON.1990.117861
