Enregistré dans:
Détails bibliographiques
Auteurs principaux: Lambe, Jason J., Chen, Feng, Stindl, Tom, Kwan, Tsz-Kit Jeffrey
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:https://arxiv.org/abs/2503.18351
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866915213616349184
author Lambe, Jason J.
Chen, Feng
Stindl, Tom
Kwan, Tsz-Kit Jeffrey
author_facet Lambe, Jason J.
Chen, Feng
Stindl, Tom
Kwan, Tsz-Kit Jeffrey
contents Terrorist activities often exhibit temporal and spatial clustering, making the multivariate Hawkes process (MHP) a useful statistical model for analysing terrorism across different geographic regions. However, terror attack data from the Global Terrorism Database is reported as total event counts in disjoint observation periods, with precise event times unknown. When the MHP is only observed discretely, the likelihood function becomes intractable, hindering likelihood-based inference. To address this, we design an unbiased estimate of the intractable likelihood function using sequential Monte Carlo (SMC) based on a representation of the unobserved event times as latent variables in a state-space model. The unbiasedness of the SMC estimate allows for its use in place of the true likelihood in a Metropolis-Hastings algorithm, from which we construct a Markov Chain Monte Carlo sample of the distribution over the parameters of the MHP. Using simulated data, we assess the performance of our method and demonstrate that it outperforms an alternative method in the literature based on mean squared error. Terrorist activity in Afghanistan and Pakistan from 2018 to 2021 is analysed based on daily count data to examine the self- and cross-excitation effects of terrorism events.
format Preprint
id arxiv_https___arxiv_org_abs_2503_18351
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Fitting multivariate Hawkes processes to interval count data with an application to terrorist activity modelling -- a particle Markov chain Monte Carlo approach
Lambe, Jason J.
Chen, Feng
Stindl, Tom
Kwan, Tsz-Kit Jeffrey
Methodology
Terrorist activities often exhibit temporal and spatial clustering, making the multivariate Hawkes process (MHP) a useful statistical model for analysing terrorism across different geographic regions. However, terror attack data from the Global Terrorism Database is reported as total event counts in disjoint observation periods, with precise event times unknown. When the MHP is only observed discretely, the likelihood function becomes intractable, hindering likelihood-based inference. To address this, we design an unbiased estimate of the intractable likelihood function using sequential Monte Carlo (SMC) based on a representation of the unobserved event times as latent variables in a state-space model. The unbiasedness of the SMC estimate allows for its use in place of the true likelihood in a Metropolis-Hastings algorithm, from which we construct a Markov Chain Monte Carlo sample of the distribution over the parameters of the MHP. Using simulated data, we assess the performance of our method and demonstrate that it outperforms an alternative method in the literature based on mean squared error. Terrorist activity in Afghanistan and Pakistan from 2018 to 2021 is analysed based on daily count data to examine the self- and cross-excitation effects of terrorism events.
title Fitting multivariate Hawkes processes to interval count data with an application to terrorist activity modelling -- a particle Markov chain Monte Carlo approach
topic Methodology
url https://arxiv.org/abs/2503.18351