安徽大学学报（自然科学版）2012，Vol.36Issue(4)：68-72,5.
基于Normalized Cut的基因表达数据聚类
Clustering of gene expression data based on Normalized Cut
摘要
Abstract
Cluster analysis with gene expression data can improve tumor diagnosis accuracy and it has a positive significance on biomedical research. This paper proposed a method based on Normalized Cut to solve the clustering problem of gene expression data. Firstly, samples were mapped to the points of high dimensional space and normalized laplacian matrix was constructed by affinity matrix and degree matrix. Then, the indicator vector reflecting category information of original samples was obtained by singular value decomposition of normalized laplacian matrix. Finally, the clustering problem was solved by using the signs of the indicator vector components. The validity of this method was proven by experiments on leukemia data and colon cancer data.关键词
聚类/指示向量/Normalized Cut/基因表达数据Key words
clustering/ indicator vector/Normalized Cut/ gene expression data分类
信息技术与安全科学引用本文复制引用
王俊生,王年,郭秀丽,唐俊..基于Normalized Cut的基因表达数据聚类[J].安徽大学学报（自然科学版）,2012,36(4):68-72,5.基金项目
国家自然科学基金资助项目(60772121) （60772121）
安徽大学“211工程”学术创新团队基金资助项目(KJTD007A) （KJTD007A）
