Logo image
Back
Document Clustering with an Augmented Nonnegative Matrix Factorization Model
Book chapter   Peer reviewed

Document Clustering with an Augmented Nonnegative Matrix Factorization Model

Zunyan Xiong, Yizhou Zang, Xingpeng Jiang, Xiaohua Hu and Xuezhi Jiang
Advances in Knowledge Discovery and Data Mining, pp 348-359
2014
DOI: https://doi.org/10.1007/978-3-319-06605-9_29

Abstract

clustering graph Laplacian nonnegative matrix factorization regularization social tagging
In this paper, we propose an augmented NMF model to investigate the latent features of documents. The augmented NMF model incorporates the original nonnegative matrix factorization and the local invariance assumption on the document clustering. In our experiment, first we compare our model to baseline algorithms with several benchmark datasets. Then the effectiveness of the proposed model is evaluated using datasets from CiteULike. The clustering results are compared against the subject categories from Web of Science for the CiteULike dataset. Experiments of clustering on both benchmark data sets and CiteULike datasets outperforms many state of the art clustering methods.

Metrics

24 Record Views
4 citations in Scopus

Details

Logo image