Abstract: Latent Semantic Analysis(LSA) lacks computation efficiency and has storage deficiencies when it is used in the large scale semantic retrieval. To solve this problem, this paper proposes a clustering-based semantic retrieval algorithm. This algorithm clusters the documents using their structural information, and applies the LSA process on those clusters to efficiently reduce the number of documents. Experimental results show that the algorithm can exponentially decrease the time of inquiring and get good retrieval accuracy.

Key words: Latent Semantic Analysis(LSA), information retrieval, vector space model, graph clustering algorithm

摘要： 潜在语义分析在进行大规模语义检索时计算效率较低、存储开销较大。针对该问题，提出一种基于聚类的潜在语义检索算法。通过文档之间的结构关系对文档进行聚类，利用簇代替文档分析潜在语义，以此减少处理文档的个数。实验结果表明，该算法能减少查询时间，且检索精确度较高。

关键词: 潜在语义分析, 信息检索, 向量空间模型, 图聚类算法

CLC Number: 

• TP301.6