Study On The Forecasting System For Dangshansu Pear Scab Based On Case-based Reasoning

Dangshansu pear has a long cultivation history and it is a good breed around the world. At Dangshansu pear planting district in the area of Yellow River valley, There are many kinds of diseases that harm Dangshansu pear, and the most serious disease is pear scab. If it is not prevented timely, this disease often endangers serious incidence and causes great loss. So the prevention against pear scab is the most important among all diseases, and accurate forecasting is the foundation to take valid prevention. In this research, based on the pear diseases' observations and weather data in the area of the Yellow River valley, we developed the forecasting system for Dangshansu pear scab, which had improved the forecasting accuracy and depended fully on computer. The system provided a new and perspective forecasting approach for the Dangshansu pear scab. The main used technologies were CBR and Fuzzy ISODATA clustering in this research. CBR is a modern expert system reasoning mode based on obtaining the problem's solution from the past experiences of solving the similar problems. Case was used to represent the past occurrence circumstance of pear scab, and knowledge acquisition and representation were natural and direct, which had overcome the bottleneck question of knowledge acquisition in the traditional RBR system. Moreover, the solved typical problem could be considered as a new case and retained in the casebase, which could get further updating. At the same time, the efficiency of the system can also get enhancement. Fuzzy ISODATA clustering is a local optimal algorithm which revises the center of the clustering unceasingly based on the extreme value of panto-function. It is based on the thought of the fuzzy partition and the iterating method. This study combined Fuzzy ISODATA clustering technology with CBR and realized the fuzzy imprecise matching reasoning in case retrieval, which emulates the reasoning of the expert experience.The forecasting system for Dangshansu pear scab was realized by VB, which is an object-oriented programming. Access was used to establish the scab casebase. The system used VB and Access to realize the maintenance of casebase and the forecasting. The interface of the system was friendly and the visual operation process was very simple to learn. Relativity statistical analysis was conducted between the real data and the forecasting data of Dangshansu pear scab, which showed the system was valid to forecast the occurrence tendency of Dangshansu pear scab.