基因表达数据分类算法及应用探讨  

Algorithm and Application Approach of Gene Expression Data Classification

作　　者：张浩[1] 叶明全[1] 汪楠[2] 

机构地区：[1]皖南医学院计算机教研室,安徽芜湖241000 [2]安庆职业技术学院电子信息系,安徽安庆246003

出　　处：《洛阳理工学院学报（自然科学版）》2014年第3期55-59,共5页Journal of Luoyang Institute of Science and Technology：Natural Science Edition

基　　金：安徽省高校自然科学研究重点项目(KJ2014A266)

摘　　要：通过机器学习方法辅助分析生物信息学中的数据,使用微阵列测试技术所获得的基因表达数据能够将任何给定条件下的基因表达模式表现出来,有利于研究人员更加深入地对众多生物过程的本质进行了解和掌握。文章对基因功能分类方法和基因表达数据的肿瘤分类进行了分析。对于基因表达数据的基因功能分类,按照功能类的隶属关系,提出基于功能树的优势因子决策和基于功能树的置信度调整准则,按照这两种标准进行基因功能树的基因功能分类算法改进。对于基因表达数据的肿瘤分类,将传统SVM算法和kNN算法两者进行结合,形成一种新型的分类算法,主要适用于肿瘤的分类。When machine learning method is used to help the analysis of bioinformatics data, the gene expression data obtained through microarray testing techniques can show the gene expression patterns under any given conditions, which contributes to the further understanding and mastering of the nature of massive biological processes for researchers. This article has analyzed the classification of gene functions and the tumor classification of gene expression data. For the former, according to the subjection relationship of function classes, dominate factor decision algorithm based on gene function tree and confidence adjustment algorithm based on gene function tree have been put forward to improve the gene function classification algorithm based on function tree. For the latter, the traditional SVM algorithm and kNN algorithm are combined to form a new classification type mainly applied to the classification of tumors.

关 键 词：基因表达数据 分类算法 应用 肿瘤分类 功能树 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]