Key Technologies For Wide Area Agricultural Information Service System Based On Ontology

With the accelerated pace of infiltration of modern information technology to the fieldof agriculture, Chinese agriculture is undergoing a transformation from traditional tomodern. China is a large agricultural country. Because of starting late of agriculturalinformation, the study of the agriculture information service is still in the developing stage.At present a variety of agricultural information service system are with low-levelautomation and few practical. There are not yet agriculture information service systemsthat can answer users’ questions automatically. There are extensive regions in China. Thecapacity of solving practical problems are lower and in uneven levels. In order toeffectively solve the organization of the agricultural information resources and provideconvenient and accurate information services for information demand, to carry outagricultural information service has become an inevitable trend of development of theagricultural information services.The agricultural information service system is studied based on the actual level ofdevelopment of rural areas in China in the thesis. It can easily implement knowledgeorganization and automatically answer users’ questions. It not only can provide real-timeand non-real-time interaction between users and experts, but also can automatically finishthe arrangement and reorganization of knowledge and realize knowledge base revision inthe natural process of question and answer. The research of the thesis, which is keytechnologies of constructing the system, has important practical significances and practicalvalues.Conclusions of the research work and innovation can be summarized as follows.(1) At present there are kinds of classification methods of agricultural issues, but theyare not suitable for constructing the large and complex agricultural information servicesystem. The classification method of agriculture issues based on modern agriculturalindustrial system is proposed in the thesis. In order to adapt the knowledge organization ofmulti-field complex agricultural issues, the agricultural knowledge is divided into threelevels, which are agricultural products, production process and the key technology. Theclassification system of ontology is analyzed and the classification ontology of agriculturalissues is described using the OWL language.(2) Facing to a variety of agricultural issues, how to classify the issues is the key toimprove the intelligence of the system and the effective service capabilities. Pre-processing and digitization of text, feature selection, weight calculation and multi-layer classificationtechnology based on support vector machine algorithm are studied in the thesis. Theontology of the multi-level classification based on agricultural issues is proposed and thevalidity of the classification method is verified finally.(3)Automatic response is an important feature of the agricultural information servicesystem. The solution mechanism of agricultural issues is studied and the similaritycalculation method, which is combined with statistical similarity and semantic similarity, isproposed and the most satisfactory answer is found out by calculating the overall similaritybetween question-and-answer pairs. When the problem can not be answered or notsatisfied with the given answer, the solution strategies and processes of submitting toexperts automatically are proposed in the thesis.(4) Knowledge base is rich or not directly affects the ability of the system services,so it is crucial to realize timely renewal and proliferation of knowledge base. The updatingprocess of the knowledge base is studied in the thesis. Solution process of the questions isrecorded to the database. The processed knowledge, which is converted into knowledge ofthe knowledge base, is used for auto-answer and auto-classification. Then informationproliferation is realized. The agricultural products name identification method based onCRF is studied. The knowledge base is updated by automatically recognizing new words.(5) The framework of the wide area agricultural information service system based onontology is constructed and the wide area agricultural information service experimentalsystem is developed based on the key technologies in this thesis. The feasibility of thetechnologies is verified through system testing.The key technologies for constructing automatic response information service systemof agricultural issues are systematic studied combined with the theory and technology ofontology. It is innovative at the aspects of classification ontology construction ofagricultural issues, automatic response and knowledge base revision.