Font Size: a A A

Research On Crime Cases Analysis Based On Spatio-temporal Data Mining

Posted on:2016-04-05Degree:MasterType:Thesis
Country:ChinaCandidate:W H WuFull Text:PDF
GTID:2346330512975330Subject:Cartography and Geographic Information System
Abstract/Summary:PDF Full Text Request
In recent years,with the further development of the public security informatization engineering(ie,the "Golden Shield Project")and deeper application of 3S technology,the public security organs has accumulated a massive spatio-temporal data of the crime cases.However,faced with the increasingly large information of crime cases and increasingly complex crime situation,the traditional methods and technology such as querying and statistics are difficult to discover the hidden association rules and trends.The circumstances of rich data and lack of knowledge seriously constrains the work of preventing and fighting against crime.Using related technology of spatio-temporal data mining,we can analyze crime cases,mining its distribution patterns and changing trends in space and time,and gain its implicit knowledge and insight.Witch has certain significance in making strategy of crime control,identifying crime patterns,optimizing the deployment of police force and planning police district and so on.Thus,we can enhance the ability of preventing and fighting against crime,improve the decision level of policing.This paper summarizes the research status of spatio-temporal data mining and analysis method of crime cases,exploratory analysis spatial and temporal characteristics of crime cases.We follow the principle of data-driven and service-oriented,use and improve existing methods and technology of spatio-temporal data mining,carry out related research.The main research work and results are as follows:(1)We summarizes the research status of spatio-temporal data model,spatio-temporal data mining and crime cases analysis.The traditional criminal geography has some limitations in big size of spatio-temporal data.This pater explore the methods and processes of spatio-temporal crime cases analysis.(2)For significant cyclical characteristics of crime cases at different time scales,this paper improve the traditional time similarity(distance)measure in clustering methods.We consider different time attribute during multiple time scales,build the equivalent spatio-temporal neighborhood and learn classical density clustering algorithm.An algorithm considering multiple time scale and equivalent spatio-temporal neighborhood based on DBSCAN(MTS-ESTN DBSCAN)was put forward.Through crime cases cluster analysis of Fuzhou in 2013,this paper shows the feasibility of the method in spatio-temporal clustering area.(3)This paper introduce the spatio-temporal autoregressive integrated moving average model(STARMA),use the crime cases data of urban area of Fuzhou during 2008 to 2010 and do empirical study on prediction analysis for crime cases and evaluate the prediction accuracy of the model.This paper verify the effectiveness of using the model to forecast the spatio-temporal crime cases.
Keywords/Search Tags:Spatio-temporal Data Mining, Spatio-temporal Clustering, Spatio-temporal Forecasting, Crime Cases
PDF Full Text Request
Related items