الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 78 |  |  |
| | | from | 78 | | |
Abstract
Case-Based Reasoning (CBR) and Rule-Based Reasoning (RBR) are two popular approaches used in decision support systems. Due to the complexities and the diversities of this domain, most medical systems become hybrid. Besides, the case adaptation process in CBR is often a challenging issue as it is traditionally carried out manually by domain experts.
Rules in RBR usually represent general knowledge, whereas cases in CBR encompass knowledge accumulated from specific (specialized) situations. Each reasoning approach has advantages and disadvantages, which are proved to be complementary in a large degree. Therefore, it is well-justified to combine rules and cases to produce effective hybrid approaches, surpassing the disadvantages of each component method.
To take advantage of the knowledge managing and reasoning capabilities in CBR and RBR, the main research objective is to present an agent-based knowledge management approach for clinical decision support systems. This approach is presented as a web service to be easily utilized. It integrates CBR and RBR, and applies the adaptation process automatically. Both adaptation rules and reasoning rules are generated from the case-base. After solving a new case, the case-base is expanded, and both adaptation and reasoning rules are updated. Besides, a prototype was implemented and experimented to diagnose breast cancer and thyroid diseases, and to classify IRIS plant type. The evaluation results showed that this approach increases the accuracy of retrieval only CBR systems, and achieves great accuracy compared to the current mammography based breast cancer diagnosis systems, thyroid diagnosis systems, and IRIS plant type classification systems.