| The prevention and cure of fruit diseases is one of the primary techniques on the fruit's production and management. The scientific forecast of fruit diseases is the foundation and precondition of the effectual prevention and cure. Case-Based Reasoning (CBR) is a modern intelligent system reasoning mode based on obtaining the problem's solution from the past experiences of solving the similar problems. It has become a technology of great prospect and potential in the field of the Artificial Intelligence and Expert System. It not only can overcome the defects in statistical model of forecast and the limitation in the traditional expert system, but also can substitute the forecasting method of fruit diseases from the experience approach to computer processing. In this paper, we firstly summarized the background, basic idea, history and development of CBR and expatiated the theory and technology of CBR. Then we discussed the character and the epidemic law of fruit diseases and analyzed the feasibility and superiority of fruit diseases forecasting system based on the Case-Based Reasoning. On the basis of them, we developed the platform of fruit diseases forecasting system based on the Case-Based Reasoning, which breaks a new approach and develops a tool for the fruit diseases forecasting. The main content and key technology on our research work is as followings: From the process of solving CBR problem, we discussed its theory and technology, such as case representation, case indexing, case retrieve, case revise, case learning and so on. According to the building of CBR system, we analyzed and demonstrated the project of fruit diseases forecasting system based on the Case-Based Reasoning and implemented every function module. According to the character of fruit diseases, we bound and organized casebase using the Database technology. We chose Template Retrieve Approach to retrieve case, which not only is rapid and high-efficiency but also can suit a great deal of case. We verified the accuracy of the forecasting system with case and SPSS statistical soft. The result showed that the correlation of the forecasting data and the real data is significant, which proves that it is feasible to forecast fruit diseases by the CBR and the result is good. The system was developed by VB that is visual and interface-friendly. It improved the practicability and the manageability. |
