欢迎访问中国科学院大学学报,今天是

中国科学院大学学报 ›› 2015, Vol. 32 ›› Issue (2): 264-272.DOI: 10.7523/j.issn.2095-6134.2015.02.017

• 计算机科学 • 上一篇    下一篇

基于多模块贝叶斯网络的恐怖袭击威胁评估

魏静1, 王菊韵2, 于华1   

  1. 1. 中国科学院大学工程管理与信息技术学院, 北京 100049;
    2. 中国传媒大学理学院, 北京 100024
  • 收稿日期:2014-03-31 修回日期:2014-05-07 发布日期:2015-03-15
  • 通讯作者: 魏静, weijingsx@163.com
  • 基金资助:

    国家重点基础研究发展计划(2011CB706900)、国家自然科学基金(70971128)和北京市自然科学基金(9102022)资助

Terrorism threat assessment with multi-module Bayesian network

WEI Jing1, WANG Juyun2, YU Hua1   

  1. 1. College of Engineering and Technology, University of Chinese Academy of Sciences, Beijing 100049, China;
    2. College of Science, Communication University of China, Beijing 100024, China
  • Received:2014-03-31 Revised:2014-05-07 Published:2015-03-15

摘要:

通过考察恐怖袭击事件可能造成的后果, 针对恐怖袭击事件评估信息的多样性、不确定性及模糊性, 提出用贝叶斯网络方法对恐怖袭击威胁进行评估, 从而为反恐决策者提供决策支持, 以减少恐怖袭击所造成的影响.针对恐怖袭击事件的复杂性, 提出多模块贝叶斯网络的恐怖袭击威胁评估模型, 并对多模块贝叶斯网络的结构学习、参数学习和推理进行研究, 提出多模块贝叶斯网络的推理算法.最后, 给出恐怖袭击威胁度的计算方法, 并对多模块贝叶斯网络的威胁评估模型进行实例分析.实例表明, 基于多模块贝叶斯网络的恐怖袭击威胁评估模型, 能有效评估恐怖袭击事件的威胁程度.

关键词: 恐怖袭击, 威胁评估, 多模块, 贝叶斯网络, 推理

Abstract:

This study intends to provide decision support for counter-terrorism according to the threat of terrorist attacks. Because of the diversity, uncertainty, and ambiguity of assessment information about terrorist attacks, Bayesian network is proposed to assess threat from the consequence of attacks. This study presents a multi-module Bayesian network threat assessment model for the complexity of the terrorist attacks, and this model combines the qualitative and quantitative assessment. We study the multi-module Bayesian network structure learning, parameter learning, and inference. Finally we compute the terrorism threat degree and conduct instance analysis. Simulation results show that this model effectively assesses the real threat degree of terrorist attacks.

Key words: terrorist attacks, threat assessment, multi-module, Bayesian network, inference

中图分类号: