Innovations in knowledge engineering , Advanced Knowledge International, Magill, South Australia. ISBN 0-9751004-0-8, 2003
Integration of various intelligent approaches, known as Hybrid Intelligent Systems (HIS), is more... more Integration of various intelligent approaches, known as Hybrid Intelligent Systems (HIS), is more common in theory and experimentation and less in applications. Specific original approaches based on combinations of con-nectionist networks and fuzzy inference systems are described and applied in the present chapter for the development of neuro-fuzzy knowledge-based models. The integration of implicit (connectionist) knowledge with (explic-it) symbolic expert systems is the main task. The modular networks para-digm is used to combine implicit and explicit knowledge in a hybrid intelli-gent system applied to toxicity prediction. Classification of toxicity correlated to the descriptors for organic compounds requires a high degree of experience from computational chemistry experts. Several approaches to generate suitable computer-based classifiers for these patterns are used to define and combine expert modules, ranging from Fuzzy Inference Systems equivalent to Quantitative Structure-Activity Relationships, classical ANN architectures, and neuro-fuzzy networks.
Bookmarks Related papers MentionsView impact
Uploads
Papers by Daniel Neagu