【作者】 杨晓蓉；

【导师】 王文生；

【作者基本信息】 中国农业科学院 ， 作物信息科学， 2011， 博士

【摘要】 随着计算机及网络技术的迅猛发展和广泛应用,农业科技信息的采集、存储、处理和传播的数量与日俱增,全国各地各部门建设了很多农业科技信息共享服务系统。但是由于涉农各种组织在建立自己的应用系统及数据存贮时,缺乏统一的规划和管理,而分别独立开发和设计各自的应用系统,因此系统分布、异构已成为现有应用环境的基本特征,导致信息资源的整合和全局应用越来越困难。而且大部分农业信息系统缺乏资源检索的语义联想能力,检索效率和准确率低下,服务模式单一,不能够根据用户的特点提供个性化信息服务。针对以上问题,在对比、分析、总结了现有的分布式信息共享关键技术的优缺点的基础上,研究确定了分布式异构农业科技信息共享中需解决的关键问题,本文设计了基于中间件法和元数据相结合的分布式农业科技信息共享框架,研究了分布式农业科技信息共享的关键技术。针对分布式农业科技信息整合问题,制定了适合农业科技信息描述的元数据标准,采用中间件自动抽取和人工描述相结合的方式实现了元数据抽取,并利用元数据副本一致性管理模型实现了元数据副本容错和负载均衡。针对农业异构数据源检索,研究了基于农业领域词典的中文分词方法和基于农业领域本体的语义扩展方法,实现了基于农业本体的查询扩展,提高了搜索的查全率和查准率。采用自动建模和人工建模相结合的方法构建用户兴趣模型,基于访问次数对用户兴趣度进行统计,群体用户的兴趣访问本体与个人用户兴趣访问本体进行聚合,实现了聚类的个性化信息资源推送。利用用户模型在本体上投影形成的个性化本体对用户查询进行针对性更强的个性化语义联想,检索出最符合用户意图的结果。最后采用以上研究的关键技术和方法,基于ASP.NET+SQL Server 2005开发环境构建了西藏分布式农业科技信息智能检索及信息共享平台,验证了本论文研究的关键技术和方法,实现了西藏地区农业科技信息的全局、统一和高效的访问。本文的创新之处主要体现在:提出了一个基于中间件法的分布式农业科技信息共享框架,可以支持农业科技信息集成化服务。在此基础上,设计了基于农业本体的异构数据库模式的自动抽取和匹配方法,提高了集成环境中农业信息智能检索的查准率和查全率;提出了基于农业本体的用户模型自动构建方法,实现了个性化信息服务。基于上述研究的技术和方法,设计开发了西藏分布式农业科技信息智能检索及信息共享平台,验证了论文所提方法的有效性。更多还原

【Abstract】 With the quick development and widespread application of computer and network technology in agricultural domain, the amount of information which is collected, saved, processed and transmitted has grown rapidly. A lot of sharing and serving systems of agricultural science and technology information are constructed by different departments throughout our country. But these systems lack a unified plan and management in the important implementation techniques and storage technologies. The system heterogeneity and dynamic distribution become basic features of the systems at present. Particularly the heterogeneity in semantics results in information sharing’s difficulty. And most agricultural information serving systems don’t support the semantic association of retrieval information. Therefore, the efficiency and correction of information serving are unsatisfactory. Most of these systems cannot also provide the personalized information service according to a user’s preference.Based on these conditions, this paper presents an architecture of the distributed agricultural science and technology information sharing based on middleware and metadata. Also the key technologies of the distributed agricultural science and technology information sharing are studied. To implement the integration of distributed agricultural science and technology information, this paper builds the metadata standards for agricultural science and technology information. Moreover, the metadata extracting method of automatic extraction based on middleware and artificial description is implemented. The metadata replica consistency controller technology is studied in detail in order to implement the metadata copy fault-tolerance and load-balanced. To implement the concept extending retrieving based on agricultural ontology, this paper studies the Chinese segmentation algorithm based on agriculture domain dictionary and semantics extending method based on agricultural ontology. To provide the personalization information services for users, this paper studies the methods of automatic modeling and artificial modeling to build a user interest model by which users’intent can be inferred. The retrieving results which satisfy users are obtained by users’personalization ontology. The information is pushed by the technology of polymerizing user groups’interest ontology and a person’s interest ontology.According to the above key technologies and methods, the Tibet intelligent retrieval and sharing platform of distributional agricultural science and technology information is developed based on ASP.NET and SQL Server 2005. The platform can provide a public and unified data access interface of different distributed data sources for users to make full use of different distributed heterogeneous data resources. In the last experiment, the intelligent retrieval and sharing platform provided proves that the architecture and methods are effective.This paper incorporates several innovations. The architecture of the distributed agricultural science and technology information sharing based on the middleware technology is presented to support the integration service of agricultural science and technology information. Based on this, the automatic extracting and matching method of heterogeneous database model based on agriculture ontology is designed to improve the efficiency and correction of agriculture information intelligent retrieving in integration environment:. The automatic user interest modelling based on agricultural ontology is proposed to implement the personalization information services. The Tibet intelligent retrieval and sharing platform of distributed agricultural science and technology information is developed to verify the efficiency of the methods proposed by this paper.更多还原