摘要
针对基因表达谱数据的高维度、低样本和连续型等特点,提出一种结合邻域互信息和自组织映射进行特征基因选取的方法.首先提出一种改进的Relief算法,对基因进行排序生成候选特征集合;然后提出基于邻域互信息的自组织映射算法对生成的候选特征基因进行聚类;最后利用提出的属性重要性系数从每一类簇中选择代表基因组成特征基因子集.实验结果表明,该方法可以快速有效地选取肿瘤特征基因,能获得较好的分类结果.
Aiming at the features of high dimension and small size of samples in gene expression data, a new method which combine neighborhood mutual information with self-organizing map for feature gene selection was proposed. Firstly, an improved Relief feature selection algorithm was proposed to sequence genes, and generate candidate feature subsets. Then, a novel Self-organizing map based on neighborhood mutual information was employed, which was used to carry out gene cluste- ring. Finally, the representative gene from each category was chose to constitute feature subset. The experiment results show that the method can select cancer informative genes promptly and effectively, and obtain better classification results.
出处
《河南师范大学学报（自然科学版）》
CAS
北大核心
2014年第1期145-150,共6页
Journal of Henan Normal University(Natural Science Edition)
基金
国家自然科学基金(60873104
61040037
61370169)
河南省科技攻关(重点)项目(112102210194)
河南省教育厅科学技术研究(重点)项目(12A520027
13A520529)
关键词
邻域互信息
RELIEF算法
自组织映射
聚类
基因选取
neighborhood mutual information
Relief algorithm
self-organizing map
clustering
gene selection
