摘要
提出了一种基于密度的聚类并行算法,在APRAM模型的分布式存储系统中,通过欧几里德距离矩阵和密度函数两次时间复杂度为O(n2)的计算,可使聚类过程的时间复杂度变为O(n),以增加一次计算的代价来降低聚类过程的时间复杂度。基于8结点的机群计算实验表明本算法能够达到较同类算法更高的并行加速比,能提高高维生物数据的聚类速度。
Put forward a clustering parallel algorithms based on the density. Use MPI under the APRAM model, passing twice computing with time complexity is O(n2) that of the Euclidean distance matrix and the density function, can make the time complexity of clustering procedure be 0 (n), reduce the time complexity of clustering through adding once computing. The experiment based on eight nodes indicates that this algorithm can attain higher parallel accelerate ratic than the same kind algorithm, raise the clustering rate of the high dimension living data.
出处
《微电子学与计算机》
CSCD
北大核心
2007年第9期130-133,共4页
Microelectronics & Computer
基金
国家自然科学基金项目(60603053)
教育部重点项目(105128)