基于演化超网络的DNA微阵列数据分类方法  

DNA microarray data classification based on evolutionary hypernetworks

作　　者：王进[1] 卢影[1] 孙开伟[1] 朱文晓[1] 赵蕊[1] 陈乔松[1] 邓欣[1] 

机构地区：[1]重庆邮电大学计算智能重庆市重点实验室,重庆400065

出　　处：《重庆邮电大学学报（自然科学版）》2014年第5期679-685,共7页Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition)

基　　金：国家自然科学基金(61203308);重庆市自然科学基金(CSTC2012jjA40034);国家大学生创新创业训练计划资助项目(201210617003)~~

摘　　要：为能够更好地从高特征维度的DNA微阵列数据中挖掘癌症相关基因,实现对恶性肿瘤的分子分型,提出了一种基于演化超网络模型的DNA微阵列数据分类方法。演化超网络是受生物网络启发而建立的一种认知学习模型,其学习过程非常适用于发掘基因间的相互作用。该方法采用信噪比进行基因选择,选择后的基因经归一化后用于演化超网络的学习和分类。通过急性白血病和结肠癌2种数据集进行实验,结果表明,演化超网络在分类精度方面与当前其他方法有较高的可比性。With the purpose of finding the cancer-related genes from the high-dimension DNA microarray data for the classi- fication of tumors, an evolutionary hypernetwork model is proposed in this paper. The evolutionary hypernetwork is a kind of cognitive learning model inspired by the biological networks and its learning process is very suitable for mining gene-gene interactions. In this paper, the gene selection is based on a signal-to-noise ratio method, and the selected normalized genes are processed by the hypernetwork through the learning and the classification phases. Empirical studies on the acute leuke- mia dataset and the colon cancer dataset demonstrate that the proposed hypernetwork classifier is high comparable with other state-of-the-art approaches in terms of classification rate of learned resuhs.

关 键 词：癌症分子分型 信噪比基因选择 演化超网络 DNA微阵列 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]