摘要
通过对慢性浅表性胃炎脾虚证与正常人、慢性浅表性胃炎脾虚证与脾胃湿热证两组胃肠粘膜配对样本的基因表达谱进行分析,在特征提取阶段分别利用W ilcoxon符号秩检验、组间和组内平方和比率(BSS/W SS)、基于相关距离的冗余分析方法,分别得到9个和12个差异基因。利用SVM作为分类器,对两组数据在此特征基因集上进行分类预测,分别达到96.67%和98.89%的预测准确率,实验结果表明该方法具有可行性和有效性。
This paper mainly aims at analysis of gastrointestinal mucosa gene expression profiles of paired samples about spleen deficiency of chronic superficial gastritis and normal people group ,and spleen deficiency of chronic superficial gastritis and piwei damp-heat group. 9 and 12 different genes are selected by using Wilcoxon signed rank test,the ratio of quadratic sum between groups and within groups (BSS/ WSS)and distance-related analysis in the process of feature selection. The forecast accuracy are 96.67% and 98.89% for two groups of data with SVM classification,which shows that the method is feasible and effective.
出处
《广西师范大学学报(自然科学版)》
CAS
北大核心
2009年第3期154-157,共4页
Journal of Guangxi Normal University:Natural Science Edition
基金
国家自然科学基金重大研究计划重点项目(90209004)
关键词
脾虚证
基因表达谱
特征提取
SVM
spleen deficiency syndrome gene expression profiles feature extraction SVM