论文部分内容阅读
化学家利用Internet通用资源搜索引擎如Yahoo、Google获取Internet资源时,检索结果常包含大量相关性较小的内容。Internet化学资源导航系统如ChemDex用人工方法收集和组织资源使资源内容质量和相关性比通用搜索引擎有所提高,但进行细致的分类仍比较困难。由中国科学院过程工程研究所建立、Internet化学化工资源导航系统ChIN的维护工具ChIN-Manag-er采用把两个内容相关的资源简介页互链来表示资源之间的密切相关关系,目前维护人员主要通用测览分类目录来确定相关资源,但这种方式在被索引的资源数量日益增大时其可用性降低。本论文开发了一种新的基于ChIN数据库检索的链接两个关系密切的相关简介页的方法。该方法针对ChIN数据库中简介页的组织特点设计了面向不同资源类型数据表的多种检索策略,这些策略侧重于被检索字段的确定;并为20余种不同类型的资源描述表建立了实现相应检索策略的检索界面。这些检索界面无缝集成到了ChIN-Manager相应的资源编辑界面中,为ChIN提供了一种快速确定被索引资源的密切相关资源的方法。
When chemists use Internet generic resource search engines such as Yahoo, Google to obtain Internet resources, the search results often contain a large number of less relevant content. Internet Chemical Resource Navigation Systems such as ChemDex use manual methods to collect and organize resources to improve the quality and relevance of resource content over common search engines, but it is still difficult to categorize them carefully. ChIN-Manag-er, a maintenance tool for the Internet chemical and chemical resource navigation system ChIN, was established by the Institute of Process Engineering of the Chinese Academy of Sciences and used to represent the close relationship between resources by interlinking the two resource-related resource brief pages. Currently, Universal browse categories to determine the relevant resources, but this approach reduces the availability of indexed resources is increasing. In this dissertation, we develop a new method to link two closely related related profile pages based on ChIN database retrieval. According to the organization features of the profile page in ChIN database, this method designs a variety of search strategies for different resource type data tables. These strategies focus on the determination of the retrieved fields, and establish the corresponding implementation for more than 20 different types of resource description tables Search Strategy Search Interface. The seamless integration of these search interfaces into the appropriate resource editing interface of ChIN-Manager provides ChIN with a quick way to determine closely related resources for indexed resources.