数字文献资源高维向量表示模型与聚类检验

High-dimensional Representation Model for Digital Literature Resource Clustering and Its Test

文献的向量表示方法对文献聚合、聚类和分类等研究是重要的。本文在向量空间模型（VSM）的基础上，通过补充文献特征此间的潜在语义相关性，提出了潜在语义向量空间模型（CLSVSM），并采用CNKI的学科分类文献为样本进行实验检验，结果显示新模型在文献聚类效果上明显好于VSM模型。

The vector representation of Literature is important for the literature aggregation, clustering and classification. The paper introduced CLSVSM based on the VSM, embedding the co-occurrence latent semantic correlation between the literatures keywords to their representative vector. So the CLSVSM is a high dimensional vector space model. Afterwards experiments were done to test the performance of new model relative to the VSM, adopt the literatures form CNKI. The results showed that CLSVSM is better than VSM model in the literature clustering testing.