An Ontology-based Fish Diseases Knowledge Acquisition And Diagnosis Reasoning Integrating System

With the prompt development of fresh water pisciculture, fish disease have happened and prevailed seriously, it has become a great obstacle of good development of aquiculture, and threaten the stabilization of aquiculture and sustainable development. The expert system of fish disease diagnose took effect in solving the conflict of the occurring of fish at frequent and expert absent. There is no common understanding, the acquisition of rule and knowledge representation is difficult, and there exist uncertainty and unreliability of inference result, when we have carefully researched the existing system. How to improve the efficiency and quality of the systems is of great value in theory and application.A fish disease diagnose knowledge concept model is provided and a methodology of building fish disease domain ontology is provided in this paper, in order to solve the problem that the difference of application aim and consumer requirement result in difference of knowledge express-, modeling methods and developing tools to solve different fields problem, often for a term we have a different understanding, lack of share understanding. These systems are separate each other, and are difficult to operate and reuse each other, bring into play low efficiency. The algorithm of transformation between XML and database is provided, it can reduce or eliminate confusion of conception and terms between the systems based-on knowledge base.On the basis of the method of machine learning, according to the idea and learning strategy of AQ algorithm, a methodology is provided which transform the problem of knowledge acquisition into the problem of combination optimization, The method or process of the model is provided based on geneticalgorithm. It can realize the function of auto- acquiring and solve the problem of bottleneck.In order to solve the problem which there exist uncertainty and unreliability of inference result, the method and theory of CBR is applied in the system. The function of feature weights is different, and in the real world they are difficult to determine. An algorithm is provided which feature weights are optimized based on genetic algorithm, the retrieval quality could be improved greatly by optimizing the feature weights. We combine RBR with CBR, and they is applied on the fish disease diagnose expert system, the methodology is availability. So the methodology in this paper could improve the overall retrieval quality and efficiency in CBR system.On the basis of the work above, we have designed and realized a fish disease diagnose reasoning system, and have tested the method and theory provided in this paper. Research indicate the system is favorable in interactive ability , intelligence and adaptability.