【作者】 刘小珠； 孙莎； 曾承； 彭智勇；

【Author】 Liu Xiaozhu~(1,2),Sun Sha~1,Zeng Cheng~1,and Peng Zhiyong~(1,3,4) 1(State Key Laboratory of Software Engineering,Wuhan University,Wuhan 430072) 2(School of Automation,Wuhan University of Technology,Wuhan 430070) 3(School of Computer,Wuhan University,Wuhan 430072) 4(Center for Studies of Information Resources,Wuhan University,Wuhan 430072)

【机构】 武汉大学软件工程国家重点实验室； 武汉理工大学自动化学院； 武汉大学计算机学院； 武汉大学信息资源研究中心；

【摘要】 倒排索引创建效率和查询效率是全文检索技术的两个重要方面.针对传统倒排索引创建方法效率低下的问题,提出了基于缓存满再写临时文件和双缓冲区相结合的索引创建机制,充分利用内存和CPU资源以加快倒排索引的创建速度;提出了查询缓存机制,以提高倒排索引的查询效率.分析及实验结果表明,提出的索引创建机制能有效地提高索引创建速度,查询缓存机制能有效地加快查询速度,提高了全文检索系统的时间和空间效率.更多还原

【Abstract】 Inverted index creation efficiency and lookup efficiency are two important aspects of full-text retrieval technology.In order to overcome the low creation efficiency of inverted index,a creation mechanism is proposed,which makes good use of memory and CPU resource.The proposed mechanism doubles buffer and writes temporary files while buffer is full.Thus,the new mechanism can speed up the index creation greatly.A query buffer mechanism is also proposed to improve lookup efficiency.Analysis and experimental results show that the proposed index creation mechanism greatly improves the speed of index creation,and the query buffer mechanism accelerates the lookup efficiency,which leads to higher efficiency of time and space in full-text retrieval system.更多还原