摘要： 提出了一种基于用户查询日志的主题式词典的构建方法,用于中文信息检索中分词。利用互信息从用户查询日志中提取短语并与通用词典相结合构建主题式词典。该词典能提高信息检索的准确率和效率,并有助于解决未登录词问题。

关键词: 主题式词典, 信息检索, 中文分词, 短语提取

Abstract: This paper proposes a novel algorithm to construct a topic-specific dictionary,based on user query log,for information retrieval.According to their mutual information,phrases are extracted from the log and they are combined with a general dictionary to construct a topic-specific dictionary.The experiment result shows that the constructed dictionary greatly improves the retrieval performance and helps to detect many out-of-vocabulary words.

中图分类号: 

• TP391.3