基于半监督聚类的真核启动子识别  

Eukaryotic Promoters Prediction Based on Semi-supervised Clustering

作　　者：赵宪佳[1] 

机构地区：[1]青岛大学国际学院,山东青岛266071

出　　处：《青岛大学学报（自然科学版）》2010年第3期42-46,共5页Journal of Qingdao University(Natural Science Edition)

基　　金：山东省自然科学基金(No.Y2008G08)

摘　　要：首先将待测试的DNA序列片段利用词项-序列矩阵进行表示,然后通过奇异值分解进行降维,最后采用全局一致性和局部一致性兼顾的半监督聚类算法对长的DNA序列片段进行测试,并与现有的几种启动子识别算法的结果进行对比。The DNA short sequences to be determined are described in the form of term-sequence matrix firstly in this paper,and then the dimension of the matrix is reduced with the method of singular value decomposition.Finally,the DNA sequences are tested by using semi-supervised clustering algorithm with global and local consistency.The comparison of the experimental result with some literatures is discussed in this paper.

关 键 词：启动子识别 半监督聚类 奇异值分解 

分 类 号：TP392[自动化与计算机技术—计算机应用技术]