【作者】 李岷； 王晓玲； 周傲英；

【Author】 LI Min,WANG Xiao-Ling,and ZHOU Ao-Ying (Department of Computer Sciense and Engineering,Fudan University,Shanghai 200433)

【机构】 复旦大学计算机科学与工程系；

【摘要】 XML已逐渐成为互联网信息的主要表示和交换工具,对XML文档的搜索引擎的研究也就越来越迫切.然而,由于XML具有结构和内容双重信息,而且结构信息复杂,XML检索技术的研究仍然面临更多挑战.现有的XML检索引擎检索效率不高,检索过程中可能出现语义错误,检索结果的表示和评估都不太符合用户的实际要求.针对现有XML检索的相关技术,提出了一种新的基于缓存和预处理技术的XML检索架构CREX,检索优化包括检索项预处理、常检索缓存等;与已有的检索引擎相比,CREX是基于语义的,充分考虑了XML文档的层次结构特点,通过优化检索项的方法达到优化检索过程的目的.此外,CREX的检索结果评估方法根据XML元素的不同特点,对中间结点和叶子结点进行分类打分.CREX具有简单和高效的特点,从一定程度上弥补了现有XML检索引擎的不足.更多还原

【Abstract】 XML has become the de facto standard for information publication and exchange on the Web. The research on XML retrieval becomes more and more important and emergent.However,due to the complicated information of XML structure,the research on XML retrieval is still challenging.The existing techniques on XML retrieval are not efficient and semantic-based.In addition,the presentation and evaluation of the search results are not applicable.Introduced in this paper are some kernel techniques including term pre-processing and classified ranking in CREX(cache-based retrieval engine of XML). Compared with the existing XML retrieval engines and techniques,CREX is semantic-based and greatly improves the efficiency of XML retrieval by optimizing query terms.Furthermore,CREX can do many improvements in ranking strategy.The ranking method ranks the query results according to the different characteristics between interior nodes and leaf node.In summary,CREX is simple and efficient and can make up the insufficiency of the existing search engines.更多还原