Pepper Nutrition Diagnosis Expert System

Great benefits have been brought by expert system's application in the fields of agriculture. As an excellent economy crop, hot-pepper has brought more and more worth to farmers. In order to increase in yield, too much fertilization and pesticide have been used. For this, the nutrition is disordered, the local ecosystem is destroyed, and the quality of hot-pepper dropped. After the domestic and abroad studying, aiming at producing requirement, and developing in our country, especially Yunnan, it's important to research hot-pepper nutrition diagnosis expert system.Considering the character of hot-pepper nutrition diagnosis and the practicality planting, the diagnosis methods have been adopted, such as plant appearance diagnosis, soil analysis diagnosis. Then we will give balancing fertilization advice for farmers, thought addition nutrient prescription fertilization, fertilizer-effect tested function etc.Knowledge acquisition is the "bottle neck" problem in the development of expert system. Considering so hard to realize the automatism acquisition of knowledge, after various resources of knowledge acquisition have been studied, the result has been used for developing hot-pepper nutrition diagnosis expert system. These resources are experts' clinic experience, schoolbooks, treatises, internet/intranets and other data. According to the characteristic of knowledge acquisition resources, the paper introduces the course of knowledge acquisition.Based on producing experience, this paper has investigated the advantage and disadvantage of the production rule knowledge representation mode. And a new mode of basing on index list is put forward which can be stored in relational database. A general model is given, too. The mode is easy to understand, implement manage, and expand, etc.After simulating experts' thinking when they are diagnosing, a mixed mode of inference, in which forward chain mainly is used, is brought forward in this paper. Firstly, forward chain will be used. To begin with known data (symptoms), the systemuses those rules with forward chain, and matches the premise of rules to the database records, and then reaches conclusion (etiology). On this basis, by the conclusion, it's easy to locate the corresponding conditions, the other diagnostic methods, and the control measures. At the same time, the inference engine based index list is designed, and the explain mechanism is introduced briefly.Interface is a communication intermedium between user and system. Based on the principles and requirement of interface designs, a designing object of interface is presented, which is concise, simple and easy to learn.All the methods, including the structure of the knowledge-base and the inference mechanism etc, have been described, and used in the development of expert system diagnosis for hot-pepper nutrition. Finally, using these results, an original system is developed. It contains diagnosis knowledge descriptions and offers the ability to diagnosis. It is proved that those methods are correct and usable.