Construction Of Potato Expert System And Discussion Of NN Knowledge Acquisition

This article introduces the present situation of the research, the structurecharacteristic, problems and development tendency of the ES. The articleelaborates the process of constructing the Shanxi intelligent ES of potatodetailedly, including the acquisition, expression and depositing of the knowledgeand the inference manner of the reasoning-machine. We used three distributionscomputations system of Browser/Web Server/Database to design systemconstruction which was open and expansible and was easy to maintain. We usedWWW Server in IIS applied by Windows 2000 Server and adopted someexploiting tools such as FrontPage 2000,Photoshop,Dreamwaver,RealProducerand Inter Dev to integrate the texts,images,animations,audios and videos toform very perfect web pages. In addition, we sufficiently used some newtechnology such as. NET,XML,ActiveX etc to strengthen the function ofsystem. On the other hand, we used the expression of knowledge based on theRBR system, mixed reasoning control strategy and the reliable model ofimprecise reasoning which resolved the complexity and imprecision in thedomain of agriculture and constructed the information management system ofpotato. The system resolved many problems of agriculture expert system such astoo huge data quantity,too large multimedia information and incertitude of data.So, the system was very popular when it was used in actual produce.With the development of the science of the technology, every walk of lifehas been affected by the development of the artificial intelligence. The expertsystem as the important off short of the expert system, it has been one of the important emphases in growing of the science. The paper begins with thedevelopment of the system, analyzing the main points of the expert system anddoing some embedded study for the method for the knowledge acquisition.The main reason about the network applying in expert system is that it cansolve the classification. The self-study of the network make the most difficultpoints the knowledge acquisition becoming the adjusting of the frame of the net,and the knowledge memorizing and distilling becoming easier. In the paper, thevirtue of the network used in knowledge acquisition has been analysesed. Somemethods of technology for the knowledge acquisition have been advanced. Inorder to improve the method of kind selection model in potato expert system,choosing the network, trying some kinds of network: Hopfield network,self-organization network, BP network. Analysising the character of thenetwork, after making many experiments, finally decided to use the BP networkfor the knowledge acquisition. Simulation for the BP network with Matlab andthe results used to the model of kind selection in potato expert system. The resultis better.