摘要
介绍潜在语义索引中半离散矩阵分解SDD(Semidiscrete Matrix Decomposition)方法的使用,改进其在存储空间方面的不足,提出了SDD+方法,并比较了奇异值分解SVD(Singular Vector Decomposition)、SDD和SDD+的性能差异。
The method of semidiserete matrix decomposistion （SDD） in latent semantic indexing is introduced. Improvements are made on storage space, and a method called SDD ＋ is proposed. Then, comparison among the performance of singular value decomposition （ SVD）, SDD and SDD ＋ is made.
出处
《计算机应用与软件》
CSCD
北大核心
2008年第7期252-253,285,共3页
Computer Applications and Software
