An IDSS Supported By Semantic Technology A Study For Swine Disease Management

In this paper, we concentrate on how to apply semantic technology(ST) and fuzzy multiple attribute decision making(MADM) in intelligence decision-making support system(IDSS) for pig disease diagnosis.As China’s advantage industry in agriculture, pig breeding is an important channel to increase farmers’ income and local finance income. In the process of pig breeding,pig disease affects the survival rate, pork quality, slaughter rate and slaughter cycle,which are directly related to the farmers’ profits. However, due to the existing problems,such as lack of veterinary in rural base and short of scientific direction for rural basic workers, the disease cannot get timely and accurate diagnosis and treatment, that makes the improvement in quality and yield of pig encounter bottleneck. Intelligence decision-making support system(IDSS) for pig disease integrating knowledge of authoritative experts which can make up for the shortage of disease experts, is a suitable approach to tackle this problem.Due to the limitations of traditional IDSS with relational database and the superiority of ST in knowledge presentation, we establish the IDSS with ST which integrates pig disease information with RDF triples. Therefore, ST-based IDSS overcomes the defects which exist in traditional IDSS because of using relational database. Furthermore, pig disease diagnosis can be described as multiple attribute decision making problem, because pig disease diagnosis is the problem that according to multiple symptom to judge disease, actually.Therefor, after analysis the problem and current solution, there are four contributions in this paper including:1) The paper proposes a framework of ST-based IDSS followed by the details of its construction for pig disease control. The framework contains three parts: semanticdatabase, pig disease diagnostic model and application program.2) A pig disease semantic database relied on pig disease semantic ontology and semantic inference is established. Pig disease ontology integrates massive, multi-source and heterogeneous pig disease experts knowledge. The paper describes the methods of ontology construction. The procedures include: knowledge extraction, knowledge pretreatment, enumeration of terms, class and class hierarchy definitions, attribute and attribute limit definitions, conversion to triples, construction rules, query verification;the principle contains: definitely clear, consistency, extendibility, preference is weak,minimal ontological commitment and be widely quoted; the development tool is Topbraid Composer;3) Inspired by the advantage of TOPSIS in MADM, pig disease diagnostic model based group and fuzzy TOPSIS is proposed. In this model, the complexity and diversity of pig disease diagnosis are adequately consider. On the basis of the knowledge database of pig disease, the paper calculate experts’ weights and integrated expert knowledge, and then utilize MADM algorithm to describe pig disease diagnosis problem. New ideas in attribute weights constructed is also provided. Finally the effectiveness of the model was validated through the experiment.4) On the basis of requirement analysis, a application of pig disease based on the pig disease semantic database and pig disease diagnostic model is developed. Myeclipse8.5 is used as integrated development environment, B/S is selected as network structure module, and MVC is chosen as development architecture. As is demonstrated by system verification, the system can give out disease diagnosis suggestions in a timely and effectively manner based on the information provided by users, such as pig’s physical conditions, symptoms or environmental information.