基于信噪比与邻域粗糙集的特征基因选择方法  

Feature Gene Selection Based on SNR and Neighborhood Rough Set

作　　者：徐久成 李涛[1,2] 孙林[1,2] 李玉惠[1,2] 

机构地区：[1]河南师范大学计算机与信息工程学院,新乡453007 [2]河南省高校计算智能与数据挖掘工程技术研究中心,新乡453007

出　　处：《数据采集与处理》2015年第5期973-981,共9页Journal of Data Acquisition and Processing

基　　金：国家自然科学基金(61370169;61402153;60873104)资助项目;河南省科技攻关重点(142102210056)资助项目;新乡市重点科技攻关计划(ZG13004)资助项目

摘　　要：鉴于传统基因选择方法会选出大量冗余基因从而导致样本预测准确率较低,提出了一种基于信噪比与邻域粗糙集的特征基因选择方法(Signal noise ration and the neighborhood rough set,SNRS)。首先采用信噪比指标获得分类能力较强的预选特征子集;然后利用邻域粗糙集约简算法对预选特征子集进行寻优;最后采用不同的分类器对特征基因子集进行分类。通过实验表明,该方法能够克服传统分类算法精度不高的缺陷,并且能够在较少的特征基因下取得较高的分类精度,验证了该方法的可行性和有效性。In view of that the traditional genetic selection method selects a large number of redundant genes,which leads to a lower sample forecast accuracy,a feature gene selection method is put forward based on the signal noise ration and the neighborhood rough set（SNRS）.Firstly,the signal-to-noise ratio（SNR）index is used to obtain the primary feature subset which have a greater impact on classification.Secondly,the rough neighborhood intensive algorithm is used to optimize the primary feature subset.Finally,feature gene subset is classified by different classifier.Experiment results show that the proposed method can get a higher classification accuracy using less feature gene than the traditional ones,which verifies the feasibility and validity of the method.

关 键 词：肿瘤基因表达谱 信噪比 邻域粗糙集 特征选择 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]