全球恐怖袭击事件的ARMA模型预测及波动周期分析

ARMA Model Forecast and Fluctuation Cycle Analysis on Global Terrorist Incidents

预测全球恐怖事件发生次数,揭示恐怖事件长期变化趋势和短期波动周期,有助于把握国际恐怖主义活动的变化态势。为此,本文以马里兰大学全球恐怖主义数据库(GTD)1999—2018年数据为样本,首先构建全球恐怖袭击事件的时间序列ARMA模型,然后开展2019—2020年全球恐怖袭击事件预测。在此基础上,利用频谱滤波(BP)方法对1970—2020年全球恐怖袭击事件的长期趋势和短期波动周期进行分析。研究主要发现:第一,2019—2020年全球恐怖袭击事情发生次数仍呈减少趋势,减少幅度预计为9.3%和5.8%左右;第二,以美国"9•11"恐怖袭击事件为分界,全球恐怖袭击事件表现出两个不同的长期变化趋势;第三,"9•11"事件以来,恐怖事件发生的次数形成了大体以7—8年为周期的短期波动,2018年以后,恐怖事件年度发生的次数及年际变化可能呈减少趋势。

Projections of the number of terrorist incidents worldwide,revealing longterm trends and short-term fluctuation cycles of global terrorist incidents help to capture the changing dynamics of global terrorist activities.Therefore,the data of the University of Maryland Global Terrorism Database(GTD)from 1999 to 2018 are used as a sample to create an ARMA(1,2)model of global terrorist incidents,and then projection is made of the global terrorist attacks from 2019 to 2020.Based on this,bandpass filter(BP)is used to analyze the long-term trends and short-term fluctuation cycles of terroristattacks from 1970 to 2020.The study mainly finds:(1)the number of global terrorist attacks from 2019—2020 may still show a decreasing trend,which is expected to be around 9.3 per cent and 5.8 per cent;(2)with the US“9/11”terrorist attacks as the watershed,global terrorist incidents have showntwo different long-term trends;(3)since“9.11”,the number of terrorist incidents has demonstrated short-term fluctuations in a seven-to-eight-year cycle,and the annual number of terrorist incidents and inter-annual changes may decrease after 2018.