采用深度置信网络的恐怖袭击事件量化分级研究
|
摘要
恐怖袭击不但会造成大量人员伤亡和财产损失,还会造成群众恐慌,对社会稳定有很大影响。旨在从数据分析角度,依据相关数据对恐怖袭击中蕴藏的信息加以分析,为防恐反恐提供有用信息。针对基于危害的恐怖袭击事件分级,根据GTD上1998-2017年数据信息进行数据提取,考虑到很多变量大面积缺失数据,首先对这些变量的重要性作简易评估并进行删减,然后对数据进行清洗和补充,最后根据特征提取和降维后的数据,分出事件对应的恐怖袭击级别。结果表明,深度学习中的深度置信网络(DBN)可以用于提取和减少预处理数据,且DBN可自动实现上述功能,无需太多人为干预。
The occurrence of terrorist attacks will not only lead to a large number of casualties and losses of property,but also cause public panic,which has a great impact on social stability and hinder people's normal work and life order. This paper aims to analyze the information contained in terrorist attacks from the perspective of data analysis according to relevant data,so as to provide useful information for counter-terrorism and counter-terrorism prevention. In view of the classification based on the harm of terrorist attacks,we first extract the data from 1998-2017 according to the GDT(Global Terrorism Databas),considering there are many variables the missing data of large area,so we first make a simple assessment of the importance of these variables,subtract them,and then clean and supplement the data. Finally,we distinguish the level of terror attack corresponding to part of the incident according to the feature extraction and dimensionality reduction data. The deep belief networks(DBN)in deep learning can be used to extract and reduce preprocessing data. DBN can do this automatically without much human intervention.
The occurrence of terrorist attacks will not only lead to a large number of casualties and losses of property,but also cause public panic,which has a great impact on social stability and hinder people's normal work and life order. This paper aims to analyze the information contained in terrorist attacks from the perspective of data analysis according to relevant data,so as to provide useful information for counter-terrorism and counter-terrorism prevention. In view of the classification based on the harm of terrorist attacks,we first extract the data from 1998-2017 according to the GDT(Global Terrorism Databas),considering there are many variables the missing data of large area,so we first make a simple assessment of the importance of these variables,subtract them,and then clean and supplement the data. Finally,we distinguish the level of terror attack corresponding to part of the incident according to the feature extraction and dimensionality reduction data. The deep belief networks(DBN)in deep learning can be used to extract and reduce preprocessing data. DBN can do this automatically without much human intervention.
引文
[1]傅子洋,徐荣贞,刘文强.基于贝叶斯网络的恐怖袭击预警模型研究[J].灾害学,2016,31(3):184-189.
[2]FRIEDMAN N,GOLDSZMIDT M,WYNER A.On the application of the bootstrap for computing confidence measures on features of induced Bayesian networks[J].AI&STAT,1999(7):189-208.
[3]赵国敏,刘茂,张青松,等.基于博弈论的地铁车站恐怖袭击风险定量研究[J].安全与环境学报,2006,6(3):47-50.
[4]MAJOR J A.Advanced techniques for modeling terrorism risk[J].Journal of Risk Finance,2002,4(1):15-24.
[5]王雷,王欣,赵秋红.基于和声搜索算法优化支持向量机的突发暴恐事件分级研究[J].管理理论,2016,28(8):125-132.
[6]LINDELAUF R H A,HUSSLASGE B G M.Cooperative game theoretic analysis of terrorist networks:the cases of Jemash Islamiyah and AIQaeda[J].European Journal of Operational Research,2013,229(1):230-238.
[7]王振,刘茂.定量风险分析在恐怖袭击风险评估中的应用[J].公共安全,2006(7):18-22.
[8]张亚军,刘宗田,周文.基于深度信念网络的事件识别[J].电子学报,2017(6):1415-1423.
[9]HINTON G E,SALAKHUTDINOV R.Reducing the dimensionality of data with neural networks[J].Science,2006,313(5768):504-507.
[10]YU D,DENG L.Deep Learning and its applications to signal and information processing[J].IEEE Signal Processing Magazine,2011,28(1):145-154.
[11]AREL I,ROSE C,KARNOWSKI T.Deep machine learning-a new frontier in artificial intelligence[J].IEEE Computational Intelligence Magazine,2010,5(4):13-18.
[12]BENGIO Y.Learning deep architectures for AI[J].Foundation and Trends in Machine Learning,2009,2(1):1-127.
[13]BENGIO Y,COURVILLE A,VINCENT P.Representation learning:a review and new perspectives[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2013,35(8):1798-1828.
[14]ANTHES G.Deep learning comes of age[J].Communications of the ACM,2013,56(6):13-15.
[15]JONES N.The learning machines[J].Nature,2014,505(7428):146-148.
[16]胡晓林,朱军.深度学习:机器学习领域的新热点[J].中国计算机学会通讯,2013,9(7):64-69.
[17]余凯,贾磊,陈雨强,等.深度学习的昨天、今天和明天[J].计算机研究与发展,2013,50(9):1799-1804.
[18]吕启,窦勇,牛新,等.基于DBN模型的遥感图像分类[J].计算机研究与发展,2014,51(9):1911-1918.
[19]陈翠平.基于深度信念网络的文本分类算法[J].计算机系统应用,2015,24(2):121-126.
[20]HINTON G E,OSINDERO S,THE Y.A fast learning algorithm for deep belief nets[J].Neural Computation,2006,18(7):1527-1554.
[21]刘亮,刘君,尹斐.基于MATLAB环境的实时仿真研究[J].微计算机信息,2006,22(7):250-252.
[2]FRIEDMAN N,GOLDSZMIDT M,WYNER A.On the application of the bootstrap for computing confidence measures on features of induced Bayesian networks[J].AI&STAT,1999(7):189-208.
[3]赵国敏,刘茂,张青松,等.基于博弈论的地铁车站恐怖袭击风险定量研究[J].安全与环境学报,2006,6(3):47-50.
[4]MAJOR J A.Advanced techniques for modeling terrorism risk[J].Journal of Risk Finance,2002,4(1):15-24.
[5]王雷,王欣,赵秋红.基于和声搜索算法优化支持向量机的突发暴恐事件分级研究[J].管理理论,2016,28(8):125-132.
[6]LINDELAUF R H A,HUSSLASGE B G M.Cooperative game theoretic analysis of terrorist networks:the cases of Jemash Islamiyah and AIQaeda[J].European Journal of Operational Research,2013,229(1):230-238.
[7]王振,刘茂.定量风险分析在恐怖袭击风险评估中的应用[J].公共安全,2006(7):18-22.
[8]张亚军,刘宗田,周文.基于深度信念网络的事件识别[J].电子学报,2017(6):1415-1423.
[9]HINTON G E,SALAKHUTDINOV R.Reducing the dimensionality of data with neural networks[J].Science,2006,313(5768):504-507.
[10]YU D,DENG L.Deep Learning and its applications to signal and information processing[J].IEEE Signal Processing Magazine,2011,28(1):145-154.
[11]AREL I,ROSE C,KARNOWSKI T.Deep machine learning-a new frontier in artificial intelligence[J].IEEE Computational Intelligence Magazine,2010,5(4):13-18.
[12]BENGIO Y.Learning deep architectures for AI[J].Foundation and Trends in Machine Learning,2009,2(1):1-127.
[13]BENGIO Y,COURVILLE A,VINCENT P.Representation learning:a review and new perspectives[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2013,35(8):1798-1828.
[14]ANTHES G.Deep learning comes of age[J].Communications of the ACM,2013,56(6):13-15.
[15]JONES N.The learning machines[J].Nature,2014,505(7428):146-148.
[16]胡晓林,朱军.深度学习:机器学习领域的新热点[J].中国计算机学会通讯,2013,9(7):64-69.
[17]余凯,贾磊,陈雨强,等.深度学习的昨天、今天和明天[J].计算机研究与发展,2013,50(9):1799-1804.
[18]吕启,窦勇,牛新,等.基于DBN模型的遥感图像分类[J].计算机研究与发展,2014,51(9):1911-1918.
[19]陈翠平.基于深度信念网络的文本分类算法[J].计算机系统应用,2015,24(2):121-126.
[20]HINTON G E,OSINDERO S,THE Y.A fast learning algorithm for deep belief nets[J].Neural Computation,2006,18(7):1527-1554.
[21]刘亮,刘君,尹斐.基于MATLAB环境的实时仿真研究[J].微计算机信息,2006,22(7):250-252.