一种基于NMF_(SC)的文本聚类方法  

Document Clustering Method Based on NMF_(SC)

作　　者：王永贵[1] 高月[1] 

机构地区：[1]辽宁工程技术大学软件学院,葫芦岛125105

出　　处：《计算机系统应用》2011年第9期78-81,156,共5页Computer Systems & Applications

摘　　要：通过分析文本的特征,提出了一种基于稀疏约束非负矩阵分解(NMFSC)的文本聚类新方法。该方法用NMFSC分解词-文本矩阵来降低特征空间的维度,并依照稀疏约束更好地控制稀疏度,然后利用簇中文本的相似性进一步细化簇。实验表明,与基于k-means的文本聚类方法和基于NMF的文本聚类方法相比,此方法具有较高的归一化互信息值(NMI),从而具有良好的聚类性能。Through analyzing the characteristics of the text, a novel text clustering approach based on Non-negative Matrix Factorization with sparseness constraint （NMFSC） is presented. The method uses NMFSC decomposing word-text matrix to reduce the dimension of the feature space, and better controls sparsity with sparseness constraint, and then further refines clusters by using the similarity of documents in clusters. Compared with text clustering method based on k-means and text clustering method based on NMF, the results of experiment show that the method has high value of the normalized mutual information, thus it has good clustering performance.

关 键 词：文本聚类 细化簇 非负矩阵分解 稀疏表示 归一化互信息值 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]