Likelihood Based Inference and Prediction in Spatio-Temporal Panel Count Models for Urban Crimes

Abstract

Standard-Nutzungsbedingungen:Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden.Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen.Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. recently developed ecient importance sampling techniques applicable to high-dimensional spatial models with sparse precision matrices. Our estimation results conrm socio-economic explanations for crime and, foremost, the broken-windows hypothesis, whereby less severe crimes in a region is a leading indicator for severe crimes. In addition to ML parameter estimates, we compute several other statistics of interest for law enforcement such as elasticities (idiosyncratic, total, short-term as well as long-term) of severe crimes w.r.t. less severe crimes, one-month-ahead out-of-sample forecasts, predictive cumulative distribution functions and validation test statistics based on these cdf's. Terms of use: Documents inJEL classication: C15; C23; C25; C51; C53; K42; R15.

Estimation and model selection in general spatial dynamic panel data models

“…, n j =1 anj ) so that each row sum of Wn is scaled to one. By these two plots, it can be seen that {y (1) it } and {y (2) it } are clearly spatially correlated.…”

“…To find a strong confirmation of the "broken-windows" phenomenon (20), we compare the differences in modeling of {y (1) i,t } without or with {y (2) i,t }. For the former one, we accordingly replace Zt by Z * t by deleting the second column of Zt .…”

“…Modeling such data is necessary but challenging. For example, urban crime count data exhibit clear spatiotemporal patterns (1), and important economic variables across space may be found that account for the concentrations of violence movement. If we can effectively model where and when crime occurs, we can launch better preventative measures.…”

“…The dataset was previously analyzed in ref. 1 by using a count model combined with a latent Gaussian spatiotemporal state process. It contains the monthly counts of crimes from January 2008 to December 2013 (72 mo) in the 138 census tracts in Pittsburgh, PA.…”

“…it in census tract i ∈[1, 138] in month t ∈[1, 72] are respectively y(j ) it = log(1 + c (j ) it ), j = 1, 2. The time plot of the averageȳ /138 is shown in Fig.…”