Research On Cloudization Method Of Agricultural Ontology Knowledge

Ontology, the clear and formal standard model of shared conceptualization, becomesthe research hotspot since it was proposed, and has got a wide-spread application in manyaspects such as knowledge engineering, information retrieval, web heterogeneousinformation processing and semantic web. In the process of building ontology, we have todeal with a lot of ontology knowledge, most of which are more or less uncertain. In orderto construct a more reasonable, objective and accurate ontology, it is necessary to doresearch on the theory and methods for dealing with the uncertainty of knowledge.In view of the shortcomings that ontology can’t perfectly express the uncertainty inagricultural, this paper focuses on three key problems in the agricultural field cloudontology modeling, including the cloudization problem of agricultural field concept,problem of classification relationship and non-classification relationship betweenagricultural field concepts. Beside, we put forward the cloudization technology andmethods for ontology knowledge, and have developed a prototype system of ontologyknowledge cloudization facing field.The main content and the research achievements are as follows:(1) The cloudization model and methods for agricultural field concepts is proposed.Through the process of statistic data of the fuzzy concepts, the fuzzy concepts are divided,and the category of the language value is determined. By using cloud model generator, itcan generate a graph which has different language value, and a cloudized concept isobtained at last;(2) The methods for the cloudization of classification relationship betweenagricultural field concepts is put forward. Through the hierarchy clustering method, theclassification relationship between concepts is extracted from the concept set, then theabove relationship is subjectively statisticed, and finally the cloudized classificationrelationships between concepts is generated through the cloud generator.(3) The cloudized methods of non-classification relationship between agriculturalfield concepts are given. Based on the association rules mining methods, non-classificationrelationships between agricultural field concepts are gained, and then the statistical data areprocessed. Finally the cloudized classification relationship between concepts is obtained byusing cloud generator.(4) A prototype system for the cloudizaton of ontology knowledge facing theagricultural field is developed. Based on the Matlab platform, a prototype system which could cloudize agricultural domain ontology knowledge is developed. And the validity ofthe used technology and methods is proved through experiments.The research results has certain theoretical value and practical significance in manyaspects, for example, the further study of the theory and method of agricultural cloudontology modeling, the building of a more accurate and objective agriculture domainontology, the establishment of the cloud model-based agricultural semantic network andknowledge grid, and the realization of fully sharing and cooperated service of theagricultural knowledge.