Figure - uploaded by Andre Python
Content may be subject to copyright.
Source publication
To this day, terrorism persists as a worldwide threat, as exemplified by the ongoing lethal attacks perpetrated by ISIS in Iraq, Syria, Al Qaeda in Yemen, and Boko Haram in Nigeria. In response, states deploy various counterterrorism policies, the costs of which could be reduced through efficient preventive measures. Statistical models able to acco...
Citations
... Even maritime piracy is highly concentrated in only a few hotspots (Marchione and Johnson 2013;Shane and Magnuson 2016). Indeed, there is already some strong evidence that terrorist attacks are tightly clustered both temporally and spatially, following concentration patterns similar to regular crime (Berrebi and Lakdawalla 2007;Townsley et al. 2008;Siebeneck et al. 2009;Medina et al. 2011;Braithwaite and Johnson 2012;Porter and White 2012;O'Loughlin and Witmer 2012;Raghavan et al. 2013;Python et al. 2016;Tench et al. 2016;Clark and Dixon 2018). ...
... Examining terrorism as an aggregated crime category makes the identification of hot-spots an almost subjective endeavor (Behlendorf et al. 2012). By examining specific types of attacks however, it may be possible to develop better analysis of events and places (Python et al. 2016). Taking such an approach is especially important given the "micro-cycle" nature of terror attacks, where specific types of attacks occur in very specific locations in "microbursts" (Behlendorf et al. 2012). ...
Objectives
This study examines the spatial characteristics of vehicular terror attacks in Israel from a “micro place” perspective at the street segment level. Utilizing data obtained from the Israel Security Agency, Israel National police, and open sources, the study analyzes the 71 vehicular attacks carried out in Israel between 2000 and 2017. In addition to examining the hot-spots at which attacks occurred, we also identify “hot routes”, estimated journey to attack routes.
Methods
We move beyond traditional approaches by calculating and comparing generalized Gini coefficients and their Lorenz curves for both the hot spots and hot routes.
Results
Tight spatial clustering in Jerusalem and the West Bank is found to be characteristic of this type of attack, which is limited by a range geographic constraints. Hot routes are identified as being highly concentrated at the street-segment level, although they are relatively less concentrated than hot-spots. Additionally, the presence of a strong distance decay function is confirmed.
Conclusions
The findings indicate that the laws of crime concentration are applicable to the case of lone terrorist vehicular attacks. The results demonstrate the utility of the methodological approach to examining specific types of terror attacks. Such approaches may be useful for informing environmental based prevention policies and strategies.
... For instance, different types of response variables may be considered, such as spatiotemporal log-Gaussian Cox processes as considered in Yuan et al. (2017). Similarly, models for data on a larger spatial scale such as global data may be fitted directly on the surface of the Earth without the need for a projection into two-dimensional space as applied in Python et al. (2016) to model global terrorism in space and time. Using the joint modelling approach these models may be extended to a multispecies or to a multievent situation. ...
As accessible and potentially vulnerable species high up in the food chain, birds are often used as indicator species to highlight changes in ecosystems. This study focuses on multiple spatially dependent relationships between a raptor (sparrowhawk), a potential prey species (house sparrow) and a sympatric species (collared doves) in space and time. We construct a complex spatiotemporal latent Gaussian model to incorporate both predator-prey and sympatric relationships, which is novel in two ways. First, different types of species interactions are represented by a shared spatiotemporal random effect, which extends existing approaches to multivariate spatial modelling through the use of a joint latent modelling approach. Second, we use a delta-gamma model to capture the semicontinuous nature of the data to model the binary and continuous sections of the response jointly. The results indicate that sparrowhawks have a localized effect on the presence of house sparrows, which could indicate that house sparrows avoid sites where sparrowhawks are present. © 2017 The Royal Statistical Society and Blackwell Publishing Ltd.
... In modeling terrorism or crime data one possibility is to use an extremely general spatio-temporal process model to capture variance not explained through the use of covariates. For example, Python et al. (2016) use a Matern class covariance function over space and an AR(1) process over time. They then use covariates to test the impact of infrastructure, population, and governance. ...
... In general, MCMC will take hours or days in order to successfully simulate from the posterior making the computational cost of fitting multiple process models extremely high. In Python et al. (2016), terrorism data was fit using a grid over the entire planet using INLA, though without self-excitation in the model. ...
... We will make the simplifying assumption that both of these are static over time. Previous statistical analysis on terrorism considered macro level covariates, such as democracy in Python et al. (2016) that differ country to country but would not change within a single country as analyzed here. Other studies considered more micro level covariates such as road networks that were found to be statistically related to terrorism in Braithwaite and Johnson (2015). ...
Spatio-temporal hierarchical modeling is an extremely attractive way to model the spread of crime or terrorism data over a given region, especially when the observations are counts and must be modeled discretely. The spatio-temporal diffusion is placed, as a matter of convenience, in the process model allowing for straightforward estimation of the diffusion parameters through Bayesian techniques. However, this method of modeling does not allow for the existence of self-excitation, or a temporal data model dependency, that has been shown to exist in criminal and terrorism data. In this manuscript we will use existing theories on how violence spreads to create models that allow for both spatio-temporal diffusion in the process model as well as temporal diffusion, or self-excitation, in the data model. We will further demonstrate how Laplace approximations similar to their use in Integrated Nested Laplace Approximation can be used to quickly and accurately conduct inference of self-exciting spatio-temporal models allowing practitioners a new way of fitting and comparing multiple process models. We will illustrate this approach by fitting a self-exciting spatio-temporal model to terrorism data in Iraq and demonstrate how choice of process model leads to differing conclusions on the existence of self-excitation in the data and differing conclusions on how violence is spreading spatio-temporally.
Over the past twenty years, research on political extremism and terrorism has become one of the fastest growing sub-fields within criminology. This rapid growth is reminiscent of the early years of criminology itself, characterized by energy, imagination and creativity but at the same time a specialization struggling to collect and analyze valid data, apply appropriate research methods and develop coherent theoretical frameworks. In this paper, we take stock of these developments by considering a basket of micro- and macro-level risk factors that have been frequently linked to the decision to engage in violent extremism. Following a review of risk factors, we consider major definitional, theoretical, data and methodological challenges and also progress made. Prior criminological research on violent extremism has focused especially on micro-level characteristics and few studies to date have integrated micro and macro determinants to explain extremist outcomes. However, with the growing availability of data, including open-source databases, paired with the application of more sophisticated statistical methods, we expect to see more robust results in the years ahead.
We introduce a new general modeling approach for multivariate discrete event data with categorical interacting marks, which we refer to as marked Bernoulli processes. In the proposed model, the probability of an event of a specific category to occur in a location may be influenced by past events at this and other locations. We do not restrict interactions to be positive or decaying over time as it is commonly adopted, allowing us to capture an arbitrary shape of influence from historical events, locations, and events of different categories. In our modeling, prior knowledge is incorporated by allowing general convex constraints on model parameters. We develop two parameter estimation procedures utilizing the constrained Least Squares (LS) and Maximum Likelihood (ML) estimation, which are solved using variational inequalities with monotone operators. We discuss different applications of our approach and illustrate the performance of proposed recovery routines on synthetic examples and a real-world police dataset.
This paper focuses on forecasting Military Action-type events by both state and non-state actors. Here we demonstrate that the dynamics of these types of events can be adequately described by a Hidden Markov Model (HMM) where the hidden states correspond to different operational regimes of an actor, and observations correspond to event frequency—and the HMM effectively predicts events with different lead times. We also demonstrate that one can enrich statistical time series-based methods that work only on historical data by exploiting predictive signals in real-time external data streams. We demonstrate the superior predictive power of the proposed models with evaluation of recent data capturing activities over two groups, ISIS and the Syrian Arab Military, two countries, Syria and Iraq, and two cities, Aleppo and Mosul. We also present an approach to converting predictions of the proposed models to real-world warnings.
This research note explains our approach to political violence forecasting employing a combination of structural and automated event data. We have created four civil war onset models using data from 1979-1999 (in sample), which we then use to predict the onset of civil war from 2000-2015 (out-of-sample). Our time horizon is one month into the future. The predictive performance of our best model as measured by its area under the receiver operating characteristic (ROC) curve is 0.881. All 24 conflict onsets between 2000-2015 occurred within countries in the top half of our risk ranking, with 79% of these onsets falling in the top 20% category, and 58% in the top 10% category. Our findings suggest that models incorporating both structural and automated event data have significantly stronger predictive power than those that rely exclusively on structural data. Conclusions and guidelines for future research are provided on the basis of our results.
