A shell environment for developing connectionist decision support systems*

Abstract

This paper presents the architecture of a neural network expert system shell. The system captures every rule as a rudimentary neural network, which is calleda network element (netel). The aim is topreserve the semantic struciure of the expert system rules, while incorporating the learning capability of neural networks into the inferencing mechanism. These netel rules are dynamically linked up to form the rule-tree during the inferencingprocess, just as a conventional expert system does. The system is also able to adjust its inference strategy according to diferent users and situations. A rule editor is provided to enable easy maintenance of the netel rules. These components are housed under a user-Fiendly interface. An application * This project is sponsored by the National Universily o/Singapore research grant RP900628.Please address aN correspondence to Tong-Seng Quah, quahtsaiss. nus.sg. expert system for USficture bonds trading is built upon this shell. The connectionist expert system has demonstrated its strength over the conventional rule-based system.

Rule-based update methods for a hybrid rule base

“…0169 Especially, there are a number of efforts combining symbolic rules and neural networks [4][5][6]8,13,16,20,22,23]. The main objective is to reduce knowledge elicitation from experts to a minimum.…”