Dynamic road crime risk prediction with urban open data

作     者：Binbin ZHOU Longbiao CHEN Fangxun ZHOU Shijian LI Sha ZHAO Gang PAN Binbin ZHOU;Longbiao CHEN;Fangxun ZHOU;Shijian LI;Sha ZHAO;Gang PAN

作者机构：College of Computer ScienceZhejiang UniversityHangzhou310027China Fujian Key Laboratory of Sensing and Computing for Smart CitiesSchool of Information Science and EngineeringXiamen UniversityXiamen361005China 

出 版 物：《Frontiers of Computer Science》 (中国计算机科学前沿（英文版）)

年 卷 期：2022年第16卷第1期

页      面：113-125页

核心收录：

学科分类：0839[工学-网络空间安全] 08[工学] 081201[工学-计算机系统结构] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：This work was partly supported by the National Natural Science Foundation of China(Grant No.61772460) Ten Thousand Talent Program of Zhejiang Province(2018R52039) 

主　　题：crime prediction road crime risk urban computing data sparsity zero-inflated negative binomial regression 

摘      要：Crime risk prediction is helpful for urban safety and citizens’life ***,existing crime studies focused on coarse-grained prediction,and usually failed to capture the dynamics of urban *** key challenge is data sparsity,since that 1)not all crimes have been recorded,and 2)crimes usually occur with low *** this paper,we propose an effective framework to predict fine-grained and dynamic crime risks in each road using heterogeneous urban ***,to address the issue of unreported crimes,we propose a cross-aggregation soft-impute(CASI)method to deal with possible unreported ***,we use a novel crime risk measurement to capture the crime dynamics from the perspective of influence propagation,taking into consideration of both time-varying and location-varying risk *** on the dynamically calculated crime risks,we design contextual features(i.e.,POI distributions,taxi mobility,demographic features)from various urban data sources,and propose a zero-inflated negative binomial regression(ZINBR)model to predict future crime risks in *** experiments using the real-world data from New York City show that our framework can accurately predict road crime risks,and outperform other baseline methods.