摘要: 查询扩展是优化信息检索的有效途径。为此,提出一种基于语义分析的查询扩展方法,利用基于互信息的共现模型分析初检文档,并将其作为部分扩展源,用模型的统计结果剪枝由语义词典WordNet生成的语义树,限制扩展范围。从初检文档和语义词典两方面选取扩展词对原查询进行扩展形成新的查询集。对返回结果进行重排序,调整前n篇文档的查准率。实验证明该方法是切实可行的。
关键词:
查询扩展,
语义树,
互信息,
文档重构
Abstract: Query expansion is an effective way to optimize information retrieval. A method for automatic query expansion based on semantic analysis is proposed. This method uses a co-occurrence model based on mutual information to analyze the retrieved documents, which is a part of the extended source, and uses the results of the analysis to prune the semantic tree generated by the semantic dictionary WordNet to limit the expansion. Extended words selected from both retrieved documents and the semantic dictionary are employed to form a new query set. The new retrieval results are re-ranked to adjust the retrieval precision. Experimental results show this method is feasible.
Key words:
query expansion,
semantic tree,
mutual information,
document reconstruction
中图分类号:
王水利, 黄广君, 霍亚格. 基于语义分析的查询扩展方法[J]. 计算机工程, 2011, 37(16): 77-79.
WANG Shui-Li, HUANG An-Jun, HE E-Ge. Query Expansion Method Based on Semantic Analysis[J]. Computer Engineering, 2011, 37(16): 77-79.