基于Cranfield体系对中文搜索引擎性能的比较研究

摘要

本文通过应用Cranfield体系——构建查询样例集合、查询信息需求分类、抓取搜索引擎结果、规定标注准则、计算性能评价指标的基本流程，对目前常用中文搜索引擎，百度、谷歌、搜狗、搜搜、人民搜索性能从定量的角度，进行较为全面的评价。其结果可供用户在选择、利用这些搜索引擎时参考。%Cranfield system through the application-to build a collection of sample query, the query classification of information needs,crawling of search engine results,provided labeling guidelines,the basic computing performance evaluation process,the most commonly used Chinese search engine,Baidu,Google,Sogou,search,people search performance from a quantitative point of view, a more comprehensive evaluation.The result for the user in the selection,use these search engines for reference.