An Improved MST Clustering Algorithm Based on Membrane Computing

An Improved MST Clustering Algorithm Based on Membrane Computing
复制标题

DOI:
10.1007/978-3-319-74521-3_1
复制
发表时间:
2017-08
期刊:
--
影响因子:
--
通讯作者:
Ping Gong;Xiyu Liu
Ping Gong;Xiyu Liu
中科院分区:
其他
文献类型:
--
作者:
Ping Gong;Xiyu Liu

文献摘要

被引文献

相似文献

MST聚类算法可以检测出边界不规则的数据簇。对于加权完全图,MST问题的可行解是非唯一的。膜计算以其分布性和最大并行性的特点,可以适当地降低图的最大生成树处理的复杂度。本文通过设计一个具体的P系统,将MST聚类算法与膜计算相结合。所设计的P系统通过收集MST问题的所有可行解,同时保留适当的边和删除冗余的重边,实现了改进的MST聚类过程。改进的MST聚类方法有效地提高了聚类质量,并通过实例证明了其可行性。
MST clustering algorithm can detect data clusters with irregular boundaries. For a weighted complete graph the feasible solutions to the MST problem is non-unique. Membrane computing known for its characteristics of distribution and maximal parallelism can properly reduce the complexity of processing a MST of a graph. This paper combines MST clustering algorithm and membrane computing by designing a specific P system. The designed P system realizes the process of an improved MST clustering by collecting all feasible solutions to the MST problem together preserving proper edges and deleting redundant heavy edges. The improved MST clustering method efficiently enhances the quality of clustering and proved to be feasible through an instance.