Research And Application Of Data Mining Technology In Global Terrorism Database

With the frequent occurrence of global terrorist incidents,the security field has gradually received the attention of all countries.Although countries have cracked down on terrorists,the global terrorist activities are still increasing.Therefore,how to use data mining methods to conduct anti-terrorism information mining with minimal cost has become more and more important.In this paper,the data on the GTD in 1998-2017 is mainly studied from three aspects:the global terrorist situation under visualization;the risk assessment model of terrorist attacks;and the spatio-temporal characteristics of terrorist attacks.In response to the terrorist situation,the GTD database data was researched using natural language and spread index.The results show that terrorist organizations prefer to use explosive means to launch terrorist attacks,while ordinary people lack security awareness and self-protection ability,which is the preferred target of attacks;extreme religiousism is still the main source of contemporary terrorist threats,and After the extinction of Al Qaeda,ISIL became a new representative terrorist organization;the motives for terrorist attacks mainly came from resource competition,religious and ethnic differences,and disputes between party and government interests;from the space spread index of terrorist activities,the 2017 terrorist attacks The incident has a tendency to spread to South America,East Asia and Oceania.It is recommended that the government departments in the region strengthen security work and take effective measures to kill terrorist activities in the cradle.A risk assessment model based on PCA and K-means cluster analysis was proposed to evaluate the risk level of terrorist attacks.Twenty risk assessment indicators were constructed by correlation analysis and chi-square test.The comprehensive score of each event was determined by solving the principal component load,and the event risk was clustered into five grades according to the score.It can be found that two of the top 10 scores were from ISIL and the targets of the attacks were soldiers or military bases,which coincided with the conclusions of the above analysis.The assessment of incident risk levels helps to visually understand the extent of the threat of terrorist attacks,as well as raise awareness of counter-terrorism in high-risk areas around the world and develop effective counter-terrorism measures.Through the global risk assessment maps of various regions,it can be found that the high-risk areas are mainly SouthAsia,the Middle East and North Africa,and South Africa.Aiming at the high-risk situation in South Asia,time-series method and spatial statistical analysis are used to study the spatio-temporal evolution characteristics.The ARIMA prediction model for establishing South Asian terrorist attacks in time predicts that the terrorist attacks in South Asia during July-October 2018 will increase significantly.Spatial autocorrelation analysis is used to study the difference of air strike activity in South Asia.The results show that the global Moran' I is greater than 0,and the P-value is less than 0.05,which is significant at the assumed level,indicating that South Asia is on the overall space.At present,Pakistan and India passed the partial spatial autocorrelation test.The spatial characteristics of the two countries are H-H aggregation mode and H-L aggregation mode respectively.Therefore,Pakistan has become a hot country in South Asia today.In addition,it should be noted that there is no L-L clustering pattern in the cluster map,indicating that all countries in South Asia have been attacked by terrorist attacks.It is recommended that South Asian countries enhance their anti-terrorism awareness during this period,abandon their former suspicions,adhere to the principles of multilateralism and peaceful coexistence,and work together for the security,stability,and prosperity of South Asia to eliminate terrorism,a global locust.