Research And Implementation Of Large Cherry Common Diseases Intelligent Diagnosis System

According to the problems of low efficiency in artificial diagnosis and weak reasoning ability in traditional expert system for large cherry diseases, this paper aims to study and realize the intelligent diagnosis system of large cherry diseases, and then to improve the diseases diagnosis ability of large cherry growers and the level of large cherry diseases prevention. On the analysis of domain knowledge of large cherry diseases, based on the improved BP neural network by genetic algorithm(GA), expert system, database and Web technology, and combining with computer network asplatform, artificial intelligence theory assupport, we focused on 10 common diseases of large cherry and established the large cherry common diseases intelligent diagnosis system based on Web using J2 EE standard three-tier B/S architecture. At the same time, the Web-based intelligent diagnosis system for cherry common diseaseshas been developed.The development of this system realized the diagnosis of cherry common diseases accurately and timely, which has provided guidance for the control of diseases occurrence rate effectively.The main research contents are the followings:(1) Building a knowledge base. Knowledge base is located in the bottom of the intelligent diagnosis system structure, and it is the guarantee of intelligent diagnosis system reliability. In order to establish a high quality of knowledge base, according to the knowledge characteristics of the 10 large cherry common diseases, based on knowledge acquisition, knowledge classification, and knowledge representation, we acquired 5 diseases diagnosis parameters of the large cherry, namely the period, part, color, shape and description of the diseases. Dynamic encoding method is adopted to generate the disease diagnosis parameter codings of large cherry automatically, and then stored them into the knowledge base. Through analysis of storge structure of the knowledge base, we build a knowledge base of large cherry disease based on the relational database, which would provide data support for the consequent establishment of the diagnosis model.(2) Establishing the diagnosis model based onimproved BP neural network. Since the traditional neural network has many weaknesses, such as slow convergence, easy to fall into local extreme points and the initial connection weights and thresholdsare lack of selection evidence, GA has been used to optimize the connection weights and thresholds. According to the design principle of BP neural network, diagnosis parameter as input variables, disease name as output variables, the network layer, the number of neurons in each layer and the training parameters were determined, so as to establish the improved BP network model based on genetic algorithm. The experimental results show that the accuracy rate of the model proposed in this paper can reach 94.86%, and compared with the traditional BP algorithm, the accuracy is 13.16% higher.(3) Realizing the intelligent diagnosis system of large cherry common diseases. On the analysis of domain knowledge and research of diagnosis model, we developed the three-tier B/S-based intelligent diagnosis system for cherry common diseases. It offers maintenance and browsing for common cherry diseases diagnosis knowledge. And users can conduct intelligent diagnosis through referring to the details of cherry diseases, and then determine the pathogen so as to take effective measures to prevent and control diseases.