首页|Application of Graph Regularized Non-negative Matrix Factorization in Characteristic Gene Selection
Application of Graph Regularized Non-negative Matrix Factorization in Characteristic Gene Selection
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
Nonnegative matrix factorization (NMF) has become a popular method and widely used in many fields, for the reason that NMF algorithm can deal with many high dimension, non-negative problems. However, in real gene expression data applications, we often have to deal with the geometric structure problems. Thus a Graph Regularized version of NMF is needed. In this paper, we propose a Graph Regularized Non-negative Matrix Factorization (GRNMF) with emphasizing graph regularized on error function to extract characteristic gene set. This method considers the samples in low-dimensional manifold which embedded in a high-dimensional ambient space, and reveals the data geometric structure embedded in the original data. Experiment results on tumor datasets and plants gene expression data demonstrate that our GRNMF model can extract more differential genes than other existing state-of-the-art methods.
School of Information Science and Engineering, Qufu Normal University, Rizhao, China
Library of Qufu Normal University, Qufu Normal University, Rizhao, China
School of Information Science and Engineering, Qufu Normal University, Rizhao, China,Bio-Computing Research Center, Shenzhen Graduate School, Harbin Institute of Technology, Shenzhen, Guangdong, China
NETL
