Supervised adaptive resonance theory and rules

• Main Author: TAN, Ah-hwee
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2000
• Subjects: Choice Function; Category Node; Confidence Factor; Category Choice; Fuzzy ARTMAP; Computer and Systems Architecture; Databases and Information Systems; OS and Networks
• Online Access: https://ink.library.smu.edu.sg/sis_research/5234; https://ink.library.smu.edu.sg/context/sis_research/article/6237/viewcontent/SART_rule.pdf
• Institution: Singapore Management University

Supervised Adaptive Resonance Theory is a family of neural networks that performs incremental supervised learning of recognition categories (pattern classes) and multidimensional maps of both binary and analog patterns. This chapter highlights that the supervised ART architecture is compatible with IF-THEN rule-based symbolic representation. Specifi­cally, the knowledge learned by a supervised ART system can be readily translated into rules for interpretation. Similarly, a priori domain knowl­edge in the form of IF-THEN rules can be converted into a supervised ART architecture. Not only does initializing networks with prior knowl­edge improve predictive accuracy and learning efficiency, the inserted symbolic knowledge can also be refined and enhanced by the supervised ART learning algorithm. By preserving symbolic rule form during learn­ing, the rules extracted from a supervised ART system can be compared directly with the originally inserted rules.