摘要
利用动态调整聚类个数的思想,在模糊C-均值聚类算法基础上引入基于多维PFS判别函数,提出一种基于多维伪F统计量的基因表达动态C-均值聚类算法。以H5N1病毒基因序列数字特征提取为例,在聚类分析过程中直接利用数字特征矩阵作为分析数据,结果表明该算法可以动态调整聚类个数,给出最佳聚类数目,从而获得较好的聚类质量。
The idea of dynamic adjustment for cluster count has been made use of in this paper, and a new dynamic C-means clustering algorithm for Genes expressed data has been proposed based on multi-dimension pseudo F-statistics. Put the numerical character of H5N1 virus gene sequence abstracted as an example,in the procedure of clustering analysis the numerical characterization matrix is directly used as analytical data source, the experiment results show that the algorithm can adjust cluster number and gain a prime number of clustering, which thus argues that this algorithm can attain better clustering quality.
出处
《计算机应用与软件》
CSCD
2009年第9期83-85,98,共4页
Computer Applications and Software
基金
辽宁省高新技术专业化人才培养重点计基金(200701A)
