浅议跨语言检索评价体系  

A Discussion on the Evaluation of Cross-Language Information Retrieval

作　　者：张艺山[1] 韩圣龙[1] 

机构地区：[1]北京大学信息管理系,北京100871

出　　处：《情报科学》2013年第7期155-160,共6页Information Science

摘　　要：本文介绍了跨语言检索系统评价体系的当前形式,对TREC、CLEF等会议的组织形式、评价方法、评价资源和评价成果进行了总结和分析。最后,论文提出在评价指标方面将DCG@n指标引入跨语言检索评价,在评价测试集方面提出利用云计算技术加强基于Web的平行语料库开发,在评价活动方面提出建立开放式的细化至系统组件的系统评价平台等观点,对跨语言检索评价体系的未来发展提出了改进意见。This article made a brief review on the status quo of the development of cross-language infor- mation retrieval and the evaluation systems. The organization and achievements of conferences that pro- motes the development of the evaluation of cross-language information retrieval, like TREC and CLEF, were discussed and analyzed. The evaluation system was composed by evaluation index, evaluation set and evaluation activities. This article gave advises to improve the future development of the evaluation system of cross-language information retrieval, such as bringing DCG@n index into cross-language information retrieval evaluation, strengthening the development of parallel corpus based on Web resources using cloud computing and building of an open system evaluation platform which focus the evaluation on the module system level.

关 键 词：跨语言信息检索 信息检索 系统评价 

分 类 号：G254.9[文化科学—图书馆学]