一种有效的基因投影聚类算法

An Efficient Algorithm of Gene Projected Clustering

针对现有基因投影聚类算法的不足,提出一种有效的基因投影聚类算法。该算法基于样本构建穷举树,根据基因间的相互作用关系,采用深度优先遍历的思想进行投影聚类,为观察疾病的成因提供了一个很好的视角。通过真实微阵列数据实验,证明了提出的算法具有较高的正确率。

To address the deficiencies of most existing gene clustering algorithms ,a novel gene projected clustering algorithm is proposed. The projected clustering is conducted based on a quick depth-first traverse on the sample enumeration tree considering the correlation among genes,which provides a new insight into the pathogenesis problem. Experimental results on real microarray data set prove high accuracy of the proposed algorithm.