Article

Analyzing behavior signatures for terrorist attack forecasting

Authors:
To read the full-text of this research, you can request a copy directly from the authors.

Abstract

In this paper we outline statistical methods used to analyze the behavior signatures that are hidden deep within data on terrorist attacks. These methods have the potential to allow military commanders to identify trends in attacks and to make informed decisions about how best to prevent future attacks. While this work focuses primarily on terrorist attacks that have occurred during Operation Iraqi Freedom, the methodology can be expanded and applied to a variety of areas. With the terrorists’ advantage of the element of surprise, analyzing their behavior may appear to be a very daunting task; however, it is demonstrated that even a collection of largely random data can often lead to insightful inferences. This paper provides a discussion of the project as a whole, the challenges faced in data collection and analysis, the methodology used, and the results and implications of the study.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

... Morris and Slocum (2012) also use the GTD and apply latent class growth analysis and general mixture modeling to identify country-level patterns of terrorism. Other studies that examine or model geographical and/or temporal patterns include: LaFree et al. (2012), Behlendorf et al. (2012), Berrebi and Lakdawalla (2007), Webb and Cutter (2009), Siebeneck et al. (2009), Brown et al. (2004, Liu et al. (2001), Inyaem et al. (2010), Reed et al. (2013), and Mohler et al. (2015). ...
Article
Full-text available
Security and intelligence agencies around the world invest considerable resources in preventing terrorist attacks, as these may cause strategic damage, national demoralization, infringement of sovereignty, and government instability. Recently, data-mining techniques have evolved to allow identification of patterns and associations in criminal data that were not apparent using traditional analysis. The aim of this paper is to illustrate how to use interpretable classification algorithms to identify subgroups ("patterns") of terrorist incidents that share common characteristics and that result in mass fatalities. This approach can produce insights far beyond those of conventional macro-level studies that use hypothesis-testing and regression models. In addition to this methodological contribution, from a practical perspective, exploring the characteristics identified in the "patterns" can lead to prevention strategies, such as alteration of the physical or systemic environment. This is in line with situational crime prevention (SCP) theory. We apply our methodology to the Global Terrorism Database (GTD). We present three examples in which terror attacks that are described by a particular pattern (set of characteristics) resulted in a high probability of mass casualties, while attacks that differ in just one of these characteristics (i.e., month of attack, geographical area targeted, or type of attack) resulted in far fewer casualties. We propose exploration of the differentiating characteristic as a means of reducing the probability of mass-fatality terrorist incidents.
... Several other publications call for more spatial pattern analysis of adverse events using Geographic Information Systems (GIS) (Berrebi and Lakdawalla, 2007;Brown et al., 2004;Johnson and Braithwaite, 2009;LaFree et al., 2012;Siebeneck et al., 2009;Webb and Cutter, 2009;Çakıt and Karwowski, 2018). Reed et al. (2011) applied a time-correlation based prediction approach to incident-based data to understand statistical patterns in terrorists' behaviors. Such models could be utilized for predicting future incidents and assigning more resources to areas that may be attacked. ...
Article
Constructing effective models for detecting, reducing, and/or preventing adverse events is very important in domains such as aviation safety, healthcare, drug administration, and war theaters. This study presents batch and data streaming models to detecting adverse events using data from a war theater context. In all the previous studies, regression models and several machine learning techniques were used for predicting continuous values in an active theater of war, and the error values reported on the test sets were large. In order to overcome the shortcoming, this study investigates the effectiveness of batch and data streaming classification algorithms in detecting or classifying adverse events given infrastructure development spending data and other variables in an active theater of war in Afghanistan. By the feature selection, the valid input variables are obtained and their indexes show that the input variables are mainly the adverse events (t-1) at the previous month, the population densities and related project investments. From the country level, fewer of the 14 project investments affect the adverse events. From the region level, some projects with higher index values, such as Security in the South Western region, Energy and Emergency Assistance in the North Eastern region, and Education in the Eastern region are mainly affecting factors. Three batch classification methods and three data streaming classification methods were assessed for their ability to detect adverse events given infrastructure development data. The study uses cost-sensitive measures to address the very unbalanced nature of the data and it applies variable reduction techniques to identify significant variables. The three batch classification algorithms are C4.5, k-nearest Neighbor, and Support Vector Machine. The three data streaming algorithms are Naïve Bayes, Hoeffding Tree, and Single Classifier Drift. In general, the performance of the cost-sensitive methods in the batch setting is comparable to those in the data stream setting. However, in the batch setting the cost matrix needs to be adjusted manually. In contrast the data stream setting allows one to adjust the models based on the analysis of the classifiers’ performance over time and changing data distribution. The Kappa values using Naïve Bayes are the highest in the three data stream algorithms in the whole country and its regions. The Naïve Bayes classifier has the best global performance. By the Kappa statistic curve, we can observe the concept drifts. In a region level, many models have a better performance including more investments related to project compared with those in a country level. In addition as data distribution becomes more balanced, the classifiers in the data stream setting outperform in terms of the overall classification rates in comparison to the classifiers in the batch setting. The results thus demonstrate the potential of data streaming algorithms to significantly outperform when the data become less unbalanced, and can be used for detecting adverse events in similar areas.
... As a prediction approach among terrorist-based data, Li, Sun, Chen, and Huang (2016) performed a prediction analysis using a neural network approach for assessing the socio-spatial dynamic of terrorist incidents in Middle East. Reed, Colley, and Aviles (2013) applied time-correlation based prediction analysis and determined trends in behaviours of terrorist groups to help decision-makers for allocating more resources and personnel to the place which are more likely to be attacked. The studies by (Çakıt et al., 2014;Çakıt & Karwowski, 2015a;Cakit & Karwowski, 2015;Çakıt & Karwowski, 2017a;Çakıt & Karwowski, 2017b (July)) used traditional statistical techniques (linear regression), data mining techniques (neural networks, adaptive neuro-fuzzy inference system, fuzzy inference system) and fuzzy C-means and subtractive clustering for predicting four categories of terrorist incidents. ...
Article
Full-text available
This study discusses a series of fuzzy overlay analysis performed within a Geographic Information System (GIS) on recent adverse events throughout the war in Afghanistan. Three types of input variables are considered in terms of number of people killed, wounded and hijacked over the period 2004–2010 in order to identify the risk level in Afghanistan using fuzzy GIS approach. To conclude, most risky areas are accumulated in the eastern region of the country and major population centres. The proposed approach could enable military decision-makers to obtain a better understanding of the socio-spatial dynamic of incidents in Middle East.
... Spatial and temporal analysis to analyze only terrorism events have been studied by several authors [11,12,13,14,15]. Some studies focused on several aspects of terrorist incidents by predicting patterns [16,17]. However, sociocultural data integrated with terrorist incidents has not been addressed. ...
Conference Paper
Full-text available
This study focused on the application of artificial neural networks (ANNs) to model the effect of infrastructure development projects on terrorism security events in Afghanistan. The dataset include adverse events and infrastructure aid activity in Afghanistan from 2001 to 2010. Several ANN models were generated and investigated for Afghanistan and its seven regions. In addition to a soft-computing approach, a multiple linear regression (MLR) analysis was also performed to evaluate whether or not the ANN approach showed superior predictive performance compared to a classical statistical approach. According to the performance comparison, the developed ANN model provided better prediction accuracy with respect to the MLR approach. The results obtained from this analysis demonstrate that ANNs can predict the occurrence of adverse events according to economic infrastructure aid activity data
... As a highly related paper to the prediction approach among terrorist based data, Reed et al. [30] aimed at demonstrating an argument that a statistical awareness of terrorists' behavior could be considered for estimating future behavior patterns. They applied a time-correlation based prediction approach and determined trends in the behavior of terrorists. ...
Article
Full-text available
Human socio-cultural behavior (HSCB) modeling has received much attention in attempts to successfully understand the effects of social and cultural factors on human behavior. The principal aim of HSCB modeling is to better organize and control the human terrain during nonconventional warfare and other missions. The intent of this study is to review and classify the literature based on the applications of different kinds of data for modeling HSCB and highlight the soft-computing methodologies suitable for HSCB event/incident data that have been applied successfully to crime data.
... In this paper, the authors highlighted several aspects of terrorist incidents by applying spatial statistics. Reed, Colley, and Aviles (2013) aimed to demonstrate a proof of concept that a statistical understanding of terrorists' behaviors could be used to predict future behavior patterns. They applied a time correlation-based prediction approach and identified trends in the behaviors of terrorists. ...
Article
Full-text available
This study investigated the relationship between adverse events and infrastructure development projects in an active theater of war using fuzzy inference systems (FIS) with the help of fuzzy clustering that directly benefits from its prediction accuracy. Fourteen developmental and economic improvement projects were selected as independent variables. These were based on allocated budgets and included a number of projects from different time periods, urban and rural population density, and total number of adverse events during the previous month. A total of four outputs reflecting the adverse events in terms of the number of people killed, wounded, or hijacked and the total number of adverse events has been estimated. The performance of each model was investigated and compared to all other models with calculated mean absolute error (MAE) values. Prediction accuracy was also tested within ±1 (difference between actual and predicted value) with values around 90%. Based on the results, it was concluded that FIS is a useful modeling technique for predicting the number of adverse events based on historical development or economic project data.
... Some authors highlighted several aspects of terrorist incidents by predicting patterns. Reed et al. (2013) aimed at demonstrating a proof of concept that a statistical understanding of terrorists' behaviors could be used to predict patterns in future behaviors. Minu, Lineesh, and John (2010) analyzed the time series of the number of terrorist attacks in the world measured on a monthly basis from 1968 to 2007. ...
Conference Paper
Full-text available
Here we report a correlation analysis aimed at determining the relationships between the occurrence of adverse events and infrastructure development in an active war theater. The total budget and total aid number of 14 project types at years t=0 (the current year), t-1 (the previous year), and t-2 (two years ago), as well as the number of adverse events (killed, wounded, and hijacked) in the previous month, and the urban and rural population density for males and females, were used to correlate the number of adverse events in year t+1. Thus, the data included a total of 89 independent variables and 4 dependent variables reflecting adverse events. All data used in this study reflect incidents and infrastructure development projects that took place between 2002 and 2010 in Afghanistan. Based on the correlation analysis findings, we conclude that most independent variables are significantly related to the dependent variables (p<0.05). The results demonstrate that past projects and population density variables are more related with dependent variables. Practitioner Summary: Analyzing and better understanding the relationships between the occurrence of adverse events and infrastructure development may provide useful information about strategies and tactics to be employed for similar scenarios in an active war theater in the future. The correlation analysis conducted in this study may serve as a starting point for the eventual development of prediction models.
Conference Paper
This study investigates the effectiveness of advanced computational intelligence techniques in detecting adverse events in Afghanistan. The study first applies feature reduction techniques to identify significant variables. Then it uses five cost-sensitive classification methods. Finally, the study reports the resulting classification accuracy rates and areas under the receiver operating characteristics charts for adverse events for each method for the entire country and its seven regions. It appears that when analysis is performed for the entire country, there is little correlation between adverse events and project types and the number of projects. However, the same type of analysis performed for each of its seven regions shows a connection between adverse events and the infrastructure budget and the number of projects allocated for the specific regions and times. Among the five classifiers, the C4.5 decision tree and k-nearest neighbor seem to be the best in terms of global performance.
Article
This paper describes a number of ways to improve on the standard method of measuring the two-point correlation, or covariance, function of large-scale structure in the universe. (1) It is shown that the statistic commonly used to estimate the covariance function has the defect that its uncertainty on large scales is limited by the uncertainty in the mean density. An alternative estimator is proposed which does not suffer from this limitation. The alternative estimator should be more accurate all on scales, but particularly on large scales where the covariance function is small. (2) Use of the sample mean biases the estimate of the covariance function by missing out variance on scales comparable to the scale of the sample. A revised estimator of the covariance function is proposed which allows missing large-scale variance to be restored. (3) It is shown from first principles how to estimate the uncertainties in the covariance function of a sample. A practical way to carry out the error analysis is suggested. (4) An approximate minimum-variance pair weighting for the calculation of the covariance function is derived, and a practical implementation is suggested. (5) The question of the calculation of background pair counts is discussed, with emphasis on the problem of the anisotropy of the redshift covariance function. Various analytic results are presented on angular integrals.