Related Experiment Videos
HYCONES II: a tool to build hybrid connectionist expert systems
B de F Leão1, A Guazzelli, E A Mendonça
1Institute of Cardiology RS, Porto Alegre, Brazil.
Proceedings. Symposium on Computer Applications in Medical Care
|January 1, 1994
Summary
HYCONES II is a new tool for building hybrid expert systems. It integrates frames with three neural network models for classification and medical diagnostics.
Area of Science:
- Artificial Intelligence
- Computational Intelligence
- Medical Informatics
Background:
- Expert systems are crucial for complex problem-solving.
- Hybrid approaches combining symbolic reasoning and neural networks offer enhanced capabilities.
- Existing tools may lack flexibility in integrating diverse neural models.
Purpose of the Study:
- To introduce HYCONES II, a novel tool for constructing hybrid connectionist expert systems.
- To present a hybrid knowledge base integrating frames with three neural network models.
- To evaluate the performance of these models in medical diagnostic classification tasks.
Main Methods:
- Development of HYCONES II, a tool for hybrid expert systems.
- Integration of frames with Combinatorial Neural Model (CNM), Fuzzy ARTMAP, and Semantic ART (SMART) neural network models.
- Representation of fuzzy medical variables within the system's symbolic component.
Main Results:
- HYCONES II enables the construction of hybrid systems for classification.
- The Semantic ART (SMART) model, a novel contribution, is introduced and evaluated.
- Comparative analysis of CNM, Fuzzy ARTMAP, and SMART models in medical diagnostic domains.
Conclusions:
- HYCONES II provides a flexible framework for hybrid expert systems.
- The integrated neural network models show potential for medical diagnostics.
- The system effectively handles fuzzy medical variables, offering avenues for future development.