Analyzing The Spatio-temporal Evolution Of Property Crime In Fuzhou

With the rapid development of economic and technological in our country,the synchronous growth property crime turns more frequent and intelligent,which means city safety has become the problems to be solved instantly.In other word,exploring the characteristics of property crime can promote our city's development in an efficient way of healthy and stable.This paper picks up four years' data of property crime including 2006?2008?2010 and 2013,and four types of crime including theft?loot?robbery and fraud,combines special places and socio-economic factors in Fuzhou,uses gravity model?Gini index?geographically weighted regression(GWR)model and combination forecasting model and so on.What's more,spatial factor?temporal factor and economic factor were selected to study the spatio-temporal evolution characteristics and driving mechanism of property crime,that is able to reflect the related properties directly.Many deficiencies were existed in traditional study,sucs as researching spatial pattern but ignoring temporal pattern,researching macro but ignoring micro.therefore,different spatial scale(county?police station area)and temporal scale(year?month?day?hour)are considered,principal component analysis and GWR model are used to excavate the spatial dependence and heterogeneity of property crime in Fuzhou.It also uses combination forecasting model to predict the next five years of property crime data in order to advance prevention recommends.The result shows that the distribution of property crime population is unbalanced and exists regional differences.Besides,the spatial structure is unstable which is reflected in the barycenter that migrates from gaishan town to cangshan town,although they are both in the cangshan region;gulou region?taijiang region and cangshan region are all belong to the crime hotspots.Property crime in Fuzhou has a strong positive correlation in the spatial distribution,and the positive correlation is enhancing year by year.The property crime population is growing from 2006 to 2013,but february owes the lowest crime data in every year.Different types of property crime have different seasonal trends.Even in 24 hours in a day,it displays different distribution.The property crime is affected by special places?socio-economic development?market circumstances and income level and so on.So this paper introduces GWR model to study the direction and degree of influence of property crime by these factors.What's more,it imports autoregressive intergrated moving average model?seasonal gray model and back propagation neural nerwork model to forecast the future crime data whose precision is high.In this paper,different spatial scale and temporal scale are considered to discuss the spaio-temporal evolution pattern of property crime in Fuzhou.It also takes advantage of visualization to combine crime analysis and information visualization,takes spatial heterogeneity and difference into account and uses GWR model while studying the driving mechanism of property crime.Finally,forecasting is researched to prevent the property crime effectively.This paper combines spatio-temporal evolution?driving mechanism and combined forecast to conduct a comprehensive research.To some extent,it enrichs the research content of property crime.