An Artificial Neural Network-based Expert System for Fruit Tree Disease and Insect Pest Diagnosis

This paper reports the development of an expert system for fruit tree disease and insect pest diagnosis based on artificial neural network (ANN) and geographic information system (GIS). A multiple knowledge acquisition approach was adopted, consisting of interview expert, questionnaire, Web-based search and literature review. The production rule was adopted as the formation of knowledge representation in the system. The reasoning process adopted a control method of depth precedence. In the prediction subsystem, the MATLAB neural network toolbox was used to predict the development tendency of fruit tree disease and insect pest. The subsystem was trained with 11 years´ meteorological information and occurrence status of fruit tree disease and insect pests in orchards of Yantai City. The ring spot, a fruit tree disease, was chosen as the research object to compare the predicted value with the actual value in this study. A GIS platform (ArcInfo) can provide the functions of spatial and temporal analysis, and was used to analyze and display the development tendency of fruit tree disease and insect pests. Preliminary results in developing a Web-based expert system for fruit tree disease and insect pest diagnosis are also summarized