基于加权动态网络的频繁模式挖掘研究  

Frequent pattern mining research based on weighted dynamic network

作　　者：肖港松[1] 陈晓云[1] 

机构地区：[1]福州大学数学与计算机科学学院,福建福州350108

出　　处：《微型机与应用》2011年第19期7-10,共4页Microcomputer & Its Applications

基　　金：国家自然科学基金(No.61070020);福建省新世纪优秀人才项目(XSJRC2007-11)

摘　　要：不同时刻的动态网络往往具有不同权重,针对加权动态网络的频繁模式挖掘,提出一种挖掘算法WGDM,它适用于加权动态社会网络、生物网络等方面的频繁模式挖掘。WGDM算法利用支持度的反单调性裁剪搜索空间,从而减少冗余候选子图,提高算法效率。通过实验测试了WGDM算法的性能,并根据中国实际股票市场网络,利用WGDM算法挖掘股票市场网络中有趣的频繁模式。The dynamic network often has different weights at different times. In order to mining the frequent pattern from the weighted dynamic network, this paper presents a mining frequent subgraph algorithm WGDM bases on weighted graph datasets. It can be applied to mining weighted dynamic network, biology network and so on. Algorithm WGDM utilizes the decrease monotony of support to reduce the search space and the number of candidate subgraph. And make a test for the algorithm WGDM. Finally use the algorithm WGDM for mining the interesting frequent pattern from the Chinese stock market network.

关 键 词：加权动态网络 加权图集 频繁子图 图挖掘 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]