Fuzzy Neural Networks and Evolving Connectionist Systems for Intelligent Decision Support

The paper presents a general framework of connectionistbased, intelligent decision support systems and its realisation with the use of fuzzy neural networks FuNNs and evolving fuzzy neural networks EFuNNs. FuNNs and EFuNNs facilitate learning from data, fuzzy rule insertion, rule extraction, and adaptation. Several applications of this framework on real problems are presented as case studies, that include classification tasks (e.g., classifying applicants for a bank loan) and on-line prediction tasks (stock index and risk assessment). The latter requires adaptive, incremental, on-line learning when the decision-making system has to work in a realtime mode. This paper suggests that fuzzy neural networks and evolving fuzzy neural networks are suitable tools to use for the implementation of the general framework.