通信学报 ›› 2019, Vol. 40 ›› Issue (9): 51-60.doi: 10.11959/j.issn.1000-436x.2019176

• 学术论文 • 上一篇    下一篇

基于节点通信行为时序的指控信息流挖掘算法

项英倬,徐正国,游凌   

  1. 盲信号处理国家重点实验室,四川 成都 610041
  • 修回日期:2019-06-11 出版日期:2019-09-25 发布日期:2019-09-28
  • 作者简介:项英倬(1990- ),男,山东东营人,盲信号处理国家重点实验室博士生,主要研究方向为数据挖掘、人工智能、大数据。|徐正国(1986- ),男,四川成都人,盲信号处理国家重点实验室工程师,主要研究方向为数据挖掘、人工智能、大数据。|游凌(1971- ),男,四川成都人,博士,盲信号处理国家重点实验室研究员、博士生导师,主要研究方向为信号分析、网络态势、数据挖掘、大数据等。
  • 基金资助:
    国家自然科学基金资助项目(61403301)

Instruction flow mining algorithm based on the temporal sequence of node communication actions

Yingzhuo XIANG,Zhengguo XU,Ling YOU   

  1. National Key Laboratory of Science and Technology on Blind Signal Processing,Chengdu 610041,China
  • Revised:2019-06-11 Online:2019-09-25 Published:2019-09-28
  • Supported by:
    The National Natural Science Foundation of China(61403301)

摘要:

针对通信网络中节点之间通信内容未知的情况,提出了一种基于节点行为时序的指控信息流挖掘算法。首先,对用户通信行为的相关性进行建模,提出了节点通信行为模型,分别对节点的背景通信和指控类通信的行为进行建模;其次,提出了 FlowMine 算法,对模型进行求解并对算法的收敛性进行了分析,该算法采用抽样迭代的思想对模型参数进行估计,能够给出参数的一个近似估计值;最后,通过模拟数据和实际数据验证并分析了FlowMine算法的有效性。实验结果表明,所提算法能够较快收敛,并能够得到可信的指控信息流。

关键词: 数据挖掘, 复杂网络, 行为时序, 信息流

Abstract:

With the situation that the content of the communication between the nodes in the network is unknown,an instruction flow mining algorithm based on the communication sequence was proposed.Firstly,by modelling the relativity of the communication actions and proposing the node communication actions model,the background communication and instruction communication actions was modelled.Moreover,FlowMine algorithm was proposed to solve such models and the convergence of the algorithm was analyzed.The algorithm estimated parameters by sampling and iteration which obtained a near optimal solution.Finally,the validity of the approach was verified by synthetic data and empirical data analysis.Experiment results show the convergence and reliable performance of FlowMine algorithm.

Key words: data mining, complex network, temporal sequence of actions, information flow

中图分类号: 

No Suggested Reading articles found!