本文综述了分布式搜索引擎的模型、结构和查询方法,并讨论了搜索引擎的评价指标。从搜索引擎的离线处理和在线处理讨论了搜索引擎的基本模块,在线查询过程速度决定了搜索引擎性能的关键因素;从分布式搜索引擎的模型上划分,搜索引擎包含四个主要子系统:网页爬虫系统、索引构建系统、检索系统和日志分析系统;倒排索引结构是以词典(dictionary)和倒排文件(inverted file)组成,分为文档编号递增排序和词频(或影响力)得分递减排序。然后讨论了当前搜索引擎典型的三类查询处理策略,并比较各自适应的条件。最后,综述评价搜索引擎的两个重要指标:查询效率和查询结果的质量,并列举定量评价公式。
This paper reviews the model,structure and search method for distributed search engine,and then discusses the evaluation of search engines. From the offline processing and online processing,the basic modules of search engine are discussed. The essential factor of search engine performance is determined by the online search processing. Divided from the distributed search engine model,the search engine consists of four main subsystems: Web crawler system,building index system,retrieval system and log analyzing system. The inverted index is divided into document ids and term frequency( or influence) sequence,which is composed of the dictionary structure and inverted file. Then the paper discusses the typically three types strategies of query processing for the current search eninge,and compares their adaptiation conditions. Finally,the two improtant indicators of evaluation of search engines are reviewed and enumerated the quantitative evaluation formula,which are query efficiently and quality of results,respectively.