城市犯罪时空同现模式的非参数检验方法

采用时空同现模式分析方法挖掘多元犯罪事件之间的关联关系,可为犯罪事件防控问题提供科学指导。现有方法依赖人为设置的频繁度阈值,应用部门若缺乏先验知识则可能导致决策错误。因此,基于非参数统计思想,提出一种面向城市犯罪的时空同现模式显著性检验方法。首先通过重建每类犯罪事件的时空分布,构建多元犯罪事件分布独立的零模型;然后根据零模型下多元犯罪事件同现频率的试验分布,判别候选时空同现模式的显著性。最后设计具有预设模式的模拟数据实验验证该方法的有效性;在多个分析尺度（时空半径）下识别S市2016年13种犯罪事件间时空同现模式,并以时空同现模式{扰乱治安,盗窃电动自行车,扒窃}为例,结合公共设施空间分布,对该模式形成机理进行深入分析。结果表明:①该方法充分顾及了单元犯罪事件自相关特征的影响,能够有效识别具有统计特性的时空同现模式;②犯罪事件时空同现模式随分析尺度的变化而存在差异;③具有相似建成环境和社会环境的犯罪事件容易形成时空同现模式。     

Using Building Permits to Monitor Disaster Recovery: A Spatio-Temporal Case Study of Coastal Mississippi Following Hurricane Katrina

The pace of disaster recovery varies considerably from one place to another. Even when places suffer impacts from the same event, recovery studies often lack the spatial and temporal resolution to fully understand such local variability in the recovery process and patterns. This paper discusses the novel use of building permits and a spatial scan statistic to identify the spatial and temporal dimensions of recovery in coastal Mississippi following Hurricane Katrina. Our work identifies significant space-time clusters of recovery activity and indicates that the amount of damage experienced and the amount of pre-event housing strongly influence the timing and location of building permit clusters. This analytical method and the use of publicly available data are valuable for a better understanding of long-term recovery processes.     

美国内布拉斯加州林肯市犯罪行为的聚类及热点分布分析

抢劫和入室盗窃犯罪行为是日常生活中最长见的违法行为,而这两种违法行为的发生有一定的规律可循。例如抢劫行为主要发生在市中心、CBD和商业中心的人群、商铺密集区域,而入室盗窃则多发生于生活社区及其周边地带。然而,想要防范这两种犯罪行为需要根据实际情况进行分析和比对,结合分析结果提出准确合理的警力部署。美国内布拉斯加州林肯市警察局针对该市抢劫和入室盗窃犯罪行为高发的情况,需要针对具体分析结果对于警力进行部署,主要针对徒步巡逻和是否需要对住户进行入户警示进行分析,根据分析结果进行具体部署。本文中使用的分析方法为紧邻层空间聚类分析(NNH)和针对犯罪的空间和时间分析法(SPAC),分别对林肯市入室行窃和抢劫的数据进行分析并尝试将两种方法进行对比以便给出更加科学的分析报告。     

Monitoring Residual Spatial Patterns using Bayesian Hierarchical Spatial Modelling for Exploring Unknown Risk Factors

This article studies Bayesian hierarchical spatial modelling that monitors the changes of residual spatial pattern (structure) of the outcome variable for exploring unknown risk factors in small‐area analysis. Spatially structured random effects (SRE) and unstructured random effects (URE) terms added to the conventional logistic regression model take into account overdispersion and residual spatial structure, which if unaccounted for could cause incorrect identification of risk factors. Mapping and/or calculating the ratio of random effects that are spatially‐structured monitor the extent of residual spatial structure. The monitoring provides insights into identification of unknown covariates that have similar spatial structures to those of SRE. Adding such covariates to the model has the potential to diminish the residual spatial structure, until possibly all or most of the spatial structure can be explained. Risk factors identified are the added covariates that have statistically significant regression coefficients. We apply the methods to the analysis of domestic burglaries in Cambridgeshire, England. Small‐area analysis of crime where data often display apparent spatial structure would particularly benefit from the methodologies. We discuss the methodologies, their relevancy in our analysis of domestic burglaries, their limitations, and possible paths for future research.     

Simulating the effects of social networks on a population’s hurricane evacuation participation

Scientists have noted that recent shifts in the earth’s climate have resulted in more extreme weather events, like stronger hurricanes. Such powerful storms disrupt societal function and result in a tremendous number of casualties, as demonstrated by recent hurricane experience in the US Planning for and facilitating evacuations of populations forecast to be impacted by hurricanes is perhaps the most effective strategy for reducing risk. A potentially important yet relatively unexplored facet of people’s evacuation decision-making involves the interpersonal communication processes that affect whether at-risk residents decide to evacuate. While previous research has suggested that word-of-mouth effects are limited, data supporting these assertions were collected prior to the widespread adoption of digital social media technologies. This paper argues that the influence of social network effects on evacuation decisions should be revisited given the potential of new social media for impacting and augmenting information dispersion through real-time interpersonal communication. Using geographic data within an agent-based model of hurricane evacuation in Bay County, Florida, we examine how various types of social networks influence participation in evacuation. It is found that strategies for encouraging evacuation should consider the social networks influencing individuals during extreme events, as it can be used to increase the number of evacuating residents.     

Population at risk: using areal interpolation and Twitter messages to create population models for burglaries and robberies

Population at risk of crime varies due to the characteristics of a population as well as the crime generator and attractor places where crime is located. This establishes different crime opportunities for different crimes. However, there are very few efforts of modeling structures that derive spatiotemporal population models to allow accurate assessment of population exposure to crime. This study develops population models to depict the spatial distribution of people who have a heightened crime risk for burglaries and robberies. The data used in the study include: Census data as source data for the existing population, Twitter geo-located data, and locations of schools as ancillary data to redistribute the source data more accurately in the space, and finally gridded population and crime data to evaluate the derived population models. To create the models, a density-weighted areal interpolation technique was used that disaggregates the source data in smaller spatial units considering the spatial distribution of the ancillary data. The models were evaluated with validation data that assess the interpolation error and spatial statistics that examine their relationship with the crime types. Our approach derived population models of a finer resolution that can assist in more precise spatial crime analyses and also provide accurate information about crime rates to the public.     

A GIS-Based Model to Determine Site Suitability of Emergency Evacuation Shelters

In recent years, the increase in the number of hurricanes and other costal hazards in the US pose a tremendous threat to the residents of coastal states. According to the National Hurricane Center, Florida is the most vulnerable coastal state to hurricanes. Mitigation policies have been formulated to reduce mortality and provide emergency services by evacuating people from the hazard zone. Many of these evacuees, particularly the elderly or lower income populations, rely on evacuation shelters for temporary housing. Because of the cost and limited use, evacuation shelters are almost exclusively dual use shelters where the primary purpose of the facility is for some other public function (e.g. school, hospital, etc.). In 2000, the estimated shortage of public shelter spaces in Florida was about 1.5 million. The purpose of this study was to rank the existing and candidate shelters (schools, colleges, churches and community centers) available in the state based on their site suitability. The research questions examined in this study include: (1) How many candidate shelters are located in physically suitable areas (e.g. not in a flood prone area, not near hazardous facilities, etc.)?; (2) How many existing shelters are located in physically un suitable areas, but in socially suitable areas (situated in areas with demand)?; (3) How many alternative existing and/or candidate shelters with high/very high physical suitability are located near physically un suitable existing shelters and thus, may be better choices for a shelter?; and (4) How many existing shelters located in physically un suitable areas are not near alternative existing and/or candidate shelters? A Geographic Information System‐based suitability model integrating Weighted Linear Combination (WLC) with a Pass/Fail screening technique was implemented for the 17 counties of Southern Florida. It was found that 48% of the existing shelters are located in physically unsuitable areas. Out of all the candidate shelters, 57% are located in physically unsuitable areas. For 15 of the existing shelters in unsuitable locations, no alternative candidate or existing shelter with medium to high physical suitability exists within 10 miles (16.1 km).     

Modeling Remote Sensing Satellite Collection Opportunity Likelihood for Hurricane Disaster Response

State and local agencies involved in emergency response to natural disasters such as hurricanes have explicitly indicated they need imagery covering the disaster area within three days of the event; and more desirably within 24 hours of the event. Airborne image collections have often been used but suffer from several problems, most noticeably the collection time (days or week) required for larger areas. The use of remote sensing satellites carrying high spatial resolution sensors has often been touted as the logical response for rapidly collecting post-disaster event imagery for emergency response. Unfortunately, satellites are maintained on fixed orbits. The repeat interval for remote sensing satellites carrying high spatial resolution sensors, even with pointable sensors, is on the order of several days, depending on the latitude for the disaster event. Fortunately, more than one satellite carries high spatial resolution imagery. This combination of requirements and restrictions may result in either a relatively high (or low) likelihood of collecting imagery within the three-day window of opportunity. This research investigated the likelihood of collecting imagery over a hurricane disaster area based on the orbital cycles of three high spatial resolution imaging satellites. Using the spatial-temporal distribution of historic hurricane landfall locations as a proxy for the probability distribution of future hurricanes by latitude, the "visibility" of each landfall location to future satellite imaging opportunities was determined. The results indicate that the likelihood of collecting imagery within one day of the event varied between 17 and 39 percent by relying on one satellite image provider. However, if either of three satellite imagery sources (i.e., Ikonos-2, Quickbird-2, and Orbview-3) could be used, then the likelihood increased to 61 percent. By relying on three satellite imagery providers there is a likelihood of between 94 and 100 percent of collecting imagery within two or three days, respectively, after the event.     

基于GIS的抢劫盗窃犯罪时空变化特征分析——以广州市为例

基于GIS方法,对广州市禁摩前和禁摩后两个时间段的抢劫盗窃犯罪案件数据,进行点模式描述和点模式分析。分析结果表明:禁摩前后广州市抢劫盗窃犯罪的案发地点在空间位置和空间范围上发生了很大变化。空间位置整体向东北方向移动,空间范围进一步扩张,由禁摩前的东西走向变为向西北和东南两个方向延伸;同时,禁摩后的空间聚集程度也有所降低。这些变化都进一步说明了,新时期下,这类犯罪呈现出了新的时空变化特征。     

A novel method for discovering spatio-temporal clusters of different sizes,shapes, and densities in the presence of noise

The discovery of spatio-temporal clusters in complex spatio-temporal data-sets has been a challenging issue in the domain of spatio-temporal data mining and knowledge discovery. In this paper, a novel spatio-temporal clustering method based on spatio-temporal shared nearest neighbors (STSNN) is proposed to detect spatio-temporal clusters of different sizes, shapes, and densities in spatio-temporal databases with a large amount of noise. The concepts of windowed distance and shared nearest neighbor are utilized to define a novel spatio-temporal density for a spatio-temporal entity with definite mathematical meanings. Then, the density-based clustering strategy is employed to uncover spatio-temporal clusters. The spatio-temporal clustering algorithm developed in this paper is easily implemented and less sensitive to density variation among spatio-temporal entities. Experiments are undertaken on several simulated data-sets to demonstrate the effectiveness and advantage of the STSNN algorithm. Also, the real-world applications on two seismic databases show that the STSNN algorithm has the ability to uncover foreshocks and aftershocks effectively.     

Diversity and burglary: Do community differences matter?

Diversity within a population has been linked to levels of both social cohesion and crime. Neighborhood crimes are the result of a complex set of factors, one of which is weak community cohesion. This article seeks to explore the impacts of diversity on burglary crime in a range of neighborhoods, using Leeds, UK as a case study. We propose a new approach to quantifying the correlates of burglary in urban areas through the use of diversity metrics. This approach is useful in unveiling the relationship between burglary and diversity in urban communities. Specifically, we employ stepwise multiple regression models to quantify the relationships between a number of neighborhood diversity variables and burglary crime rates. The results of the analyses show that the variables that represent diversity were more significant when regressed against burglary crime rates than standard socio‐demographic data traditionally used in crime studies, which do not generally use diversity variables. The findings of this study highlight the importance of neighborhood cohesion in the crime system, and the key place for diversity statistics in quantifying the relationships between neighborhood diversities and burglary. The study highlights the importance of policy planning aimed at encouraging community building in promoting neighborhood safety.     

Interaction of Hurricane Katrina With Optically Complex Water in the Gulf of Mexico: Interpretation Using Satellite-Derived Inherent Optical Properties and Chlorophyll Concentration

When Hurricane Katrina passed over southern Florida, Florida Bay and the West Florida Shelf, and into the Gulf of Mexico, empirically derived chl a increases were observed in the Tortugas Gyre circulation feature, and in adjacent waters. Analysis of the empirically derived chl a increase within the gyre has been primarily attributed to initiation of a phytoplankton bloom promoted by nutrients upwelled by Katrina's winds. Detailed analysis of inherent optical properties derived from remotely sensed radiances, however, indicated the interaction of Katrina with shallow coastal and shelf waters likely entrained waters with higher concentrations of chromophoric dissolved organic matter (CDOM) into the gyre circulation, augmenting the chl a signal. Storm-induced upwelling would also transport optically active CDOM to the surface. Increases in empirically derived chl a in the Florida coastal waters influenced by Katrina's winds were therefore partly due to increased absorption by CDOM. This analysis indicates that elevated empirically derived chl a in hurricane-influenced waters should not be unambiguously attributed to increased phytoplankton productivity, particularly in an optically complex coastal environment.     

Visualising Crime Clusters in a Space‐time Cube: An Exploratory Data‐analysis Approach Using Space‐time Kernel Density Estimation and Scan Statistics

For an effective interpretation of spatio‐temporal patterns of crime clusters/hotspots, we explore the possibility of three‐dimensional mapping of crime events in a space‐time cube with the aid of space‐time variants of kernel density estimation and scan statistics. Using the crime occurrence dataset of snatch‐and‐run offences in Kyoto City from 2003 to 2004, we confirm that the proposed methodology enables simultaneous visualisation of the geographical extent and duration of crime clusters, by which stable and transient space‐time crime clusters can be intuitively differentiated. Also, the combined use of the two statistical techniques revealed temporal inter‐cluster associations showing that transient clusters alternatively appeared in a pair of hotspot regions, suggesting a new type of “displacement” phenomenon of crime. Highlighting the complementary aspects of the two space‐time statistical approaches, we conclude that combining these approaches in a space‐time cube display is particularly valuable for a spatio‐temporal exploratory data analysis of clusters to extract new knowledge of crime epidemiology from a data set of space‐time crime events.     

一种局部多项式时空地理加权回归方法

基于加权最小二乘估计的时空地理加权回归方法,在随机项方差相同且最小的假设条件下估计回归参数和拟合值,由于没有考虑时空分析中异方差影响而导致估计结果存在一定偏差。局部多项式估计是一种消除异方差影响的非参数估计方法。本文在局部多项式估计原理基础上,提出了局部多项式时空地理加权回归方法。它是采用三元一阶泰勒级数展开式重构时空回归系数和自变量矩阵,进而建立满足高斯-马尔可夫独立同分布假定要求的新模型,利用新模型回归系数估计值、拟合值以及新模型与原模型的关系,可得到原模型回归系数估计值和拟合值。本文采用模拟数据和真实数据进行试验,以GTWR与局部线性地理加权回归作为对比方法,从方法适用性、整体估计效果、回归系数估计偏差和拟合优度、整体估计偏差等方面分析了LPGTWR方法性能,有效证明了LPGTWR方法能消除异方差影响提升估计精度。     

一种基于重排检验的时空聚类方法

融合时空邻近与专题属性相似的时空聚类是挖掘地理现象时空演化规律的重要手段。现有方法需要的聚类参数许多难以获取,影响了聚类方法的可操作性与聚类结果的可靠性。提出一种基于重排检验的时空聚类方法。首先,通过重排检验发现时空数据集中的均质子区域;进而,采用均方误差准则合并均质子区域内的时空实体生成时空簇,并通过簇内重排检验自动识别聚类合并的终止条件;最后,借助时空拓扑关系在保证结果精度的前提下发展一种快速重排检验的方法,提高了聚类方法的运行效率。通过实验和比较发现,该方法一方面可以发现不同形状、大小的时空簇,聚类质量优于经典的ST-DBSCAN方法;另一方面聚类过程中人为设置参数的主观性显著降低,提高了聚类方法的可操作性。     

基于Web新闻的案(事)件抽取与时空分析系统

新闻,自古以来便是人们了解社会动态的重要途径,大数据时代,由于Web新闻自身所具有的客观性和真实性,其蕴含的数据价值凸显。针对新闻网站中案(事)件信息丰富、易采集等优点,研究开发一套基于Web新闻的案(事)件抽取与时空分析系统,抓取各个新闻网站对发生于福州的案(事)件相关信息的报道,对新闻信息进行判别清洗与解析,采用支持向量机进行案(事)件类别分类,多类别分类精度达75%,抽取经分类处理之后的案(事)件文本中的案(事)件时空信息并进行时空分析,以毒品案(事)件为例,将解析结果与公安毒品案(事)件分别做核密度估计,结果表明,福州毒品事件集中发生于茶园派出所和象园派出所等辖区。该系统有利于分析福州社会动态,也为公安部门提供了信息辅助。     

Comparing spatially varying coefficient models: a case study examining violent crime rates and their relationships to alcohol outlets and illegal drug arrests

In this paper, we compare and contrast a Bayesian spatially varying coefficient process (SVCP) model with a geographically weighted regression (GWR) model for the estimation of the potentially spatially varying regression effects of alcohol outlets and illegal drug activity on violent crime in Houston, Texas. In addition, we focus on the inherent coefficient shrinkage properties of the Bayesian SVCP model as a way to address increased coefficient variance that follows from collinearity in GWR models. We outline the advantages of the Bayesian model in terms of reducing inflated coefficient variance, enhanced model flexibility, and more formal measuring of model uncertainty for prediction. We find spatially varying effects for alcohol outlets and drug violations, but the amount of variation depends on the type of model used. For the Bayesian model, this variation is controllable through the amount of prior influence placed on the variance of the coefficients. For example, the spatial pattern of coefficients is similar for the GWR and Bayesian models when a relatively large prior variance is used in the Bayesian model.      

顾及时空"点-线"关联特征的出行轨迹描述模型

关联出行研究是城市规划、交通出行、传染病防控、犯罪侦查等领域的研究热点,尤其当出行个体带有接头、等待等主观意图的前提下,如何有效识别其关联行为,是当前时空认知领域的难点。本文以犯罪时空轨迹为研究对象,首先分析了现有出行轨迹描述方法的不足,兼顾了出行轨迹的细节点特征和完整线特征,提出了顾及"点-线"特征的出行轨迹描述模型。然后,在此基础上,基于时空拓扑关系提出关联出行的轨迹判别方法,能描述"接头、等待、共处、同行"4种基本的关联出行轨迹模式,并可进一步区分19种不同的关联出行子类型。最后,通过试验对比,验证了该方法的有效性。     

顾及时空信息的暴雨精细化大数据平台构建方法

暴雨是城市灾害中最常见的一种,具有降水时间短、强度大的特点,很难做到精确预报。以时空信息为基础框架,建立了暴雨精细化工程综合时空数据库,研发了暴雨精细化工程基础数据平台,提供了暴雨专题数据可视化展示、暴雨灾害预测分析、灾害模拟、撤离避灾路线制定等功能。通过对暴雨的时空分布进行实时展示,可获得基于监测数据的较为准确的预测结果。     

点集数据不规则形状时空异常聚类模式挖掘研究

传统扫描统计方法在进行时空异常聚类模式挖掘时,受扫描窗口形状的限制,不能准确地获取聚类区域形状。提出一种改进的不规则形状时空异常聚类模式挖掘方法stAntScan。新方法基于26方位时空邻近单元格构建时空邻接矩阵,再对蚁群最优化扫描统计方法进行改进,使其能适应三维大数据量的时空区域扫描。模拟数据和真实微博签到数据的实验证明,stAntScan能有效地识别时空范围内的不规则形状异常聚类,并且准确性较经典的SaTScan方法高。