Method Of Crime Prediction Based On Spatial-temporal Charactieristics

In recent years,many types of criminal cases are spreading everywhere and threatening more to society.As one of the most common types,burglary is threatening the security of society and life of the mass for its massive,wide range of hazards,strong concealment,and more probability of inducing other crime cases.It is urgently needed for scientist and government to study when and where one burglary will occur based on the spatial-temporal characteristics.In order to predict burglary crime,one forecast model were proposed based on the spatial-temporal pattern mined.This research is significationly different from others as follows:1.The scene was constructed to analysis the spatial-temporal characteristics and distribution pattern of burglary.By this way,the spatial-temporal mechanism and distribution pattern quantified will be linked together.This method will not only inherit the advantage of geographical analysis,but also is conducive to the discovery of spatial-temporal distribution pattern.2.In view of the phenomenon that most of the existing researches are analyzed on a single scale,the related factors of crime are analyzed on the spatial scales of cell,neighborhood,community,street and district separately in this paper.Four conclusions are derived:?The spatial distribution of the crime is associated with different geographic factors on different spatial scales.?Different environmental criminology are needed to analyze criminal behavior in different scales.?Accuracy of burglary explained by geographic factors varies on different scales;?Compared with the distribution explained by geographic factors on one single scale.Accuracy of burglary distribution explained on multiple scales is higher.3.A spatial-temporla clustering method with none parameters is proposed,which is more accurate comparing with other coclustering algorithm.The distribution varies in different spatial-temporal unit correlated with different geographical factors.Among which,crime transfers along "Central North-North-Southeast" from Monday to Sunday.The extension and frequency of crime appear to decrease with decreasing temperature,and gradually increase with the increase of temperature monthly from summer.In summer,crime occurs in a wider range with higher frequency.At last,the distribution pattern of crime in different spatial-temporal units,and the geographical factors associated with which is analysed.4.Burglary cases are aggregated in time and space.Based on this finding,we propose one method to extract the spatial-temporal distribution pattern taking into account the spatial differences,three conclution can be draw:Conclusion one:burglary follows a power-lower distribution in space.Conclusion two:burglary follows a passion distribution in time while the distribution parameters are propotional to the number of burglary cases within the scene.Conclusion three:burglary cases are aggregated in space rather than time,which means most burglary cases take place within few scenes,and vice versa.The result is crucial for hot spot analysis and quantitive of the distribution pattern.5.Based on the spatial-temporal distribution of burglary,we propose a spatial-temporal scanning method to quantify the near-repeat pattern.Three conclusions can be draw:Conclusion one:the probability of near-repeat is proposional to the attractiveness of scenes.Conclusion two:the probability of burglary is affected by the density of burglary taken place in the past.We explain the results using invironment criminology theory at last.6.We propose a forcast model based on the distribution pattern and quantization parameters of burglary.Fisrt of all,we construct the model using the theory of attraction-repulsion of molecules:the two molecules have a strong attraction in the far distance,and repulsion can ignore.But the repulsive force increase much more than attraction while the distance reducing.Burglary also has the attractive and repulsive forces.That means,every crime would "attract" more burglary cases,and too much burglary will "reject" other burglary cases.Secondly,we analysis the temporal,spatial patterns and response of the burglary taken place past after fitting parameters of the model.The model is found to reflect the pattern of burglary extracted before.Thirdly,we test the applicability of our model in different space and time.The performace of our model is better within the scene attract more burglary cases.7.A crime transmission network based on the impact strength of nodes is constructed using the method proposed.The characteristic parameters of complex networks are introduced.The concept of degree,average degree,clustering coefficient and analysis of criminal network are carried out.The results show that degree of the nodes is closely related to the crime rate future,which can be used in crime forecasting;Distribution of node's degree has no scaling property.Burglary can also occur within areas rarely be infringed.And the degree of the node is closely related to the crime rate future.Therefore,even if the crime rate is relatively low,the district should also pay attention to the changes of node's degree;Coefficient of crime aggregation has predictability to the future crime rate.Higher aggregation coefficient means the state of crime may change in the future.The spatial-temporal characteristic of important nodes in crime prediction was analyzed at last.This dissertation took full advantage of multi-resolution characteristics of burglary,quantified the spatial-temporal pattern,and constructed a forcast model at last.The conclusions will not only deepen our understanding of the spatial-temporal pattern of burglary,enrich the method of environment criminology,but also is helpful for optimizing the deployment of police,maintening the social security and building a harmonious society.