Society has been constantly changing due to the rapid growth of cities and the rise of economic levels. Along with this social change, crime which one of the threats to society, continues to increase. The crime has become more severe over time. As the crime diversified and the crime incidence increa...
Society has been constantly changing due to the rapid growth of cities and the rise of economic levels. Along with this social change, crime which one of the threats to society, continues to increase. The crime has become more severe over time. As the crime diversified and the crime incidence increased, the need for a predictive crime prevention strategy attracted attention. In the case of foreign countries, most countries provide accurate information on crime data to predict Crime Risky Areas and minimize crime damage. However, the accurate information has not been provided to citizen in korea due to the leakage of personal information. Furthermore, it provides only crime occurrence area information of the global level. For this reason, Such information was less effective to korea citizen in predicting and preparing Crime Risky Areas. Therefore, in order to improve the efficiency of citizen's own crime prevention in Korea, it is necessary to prepare a plan to minimize crime damage without actual crime data. This study explored to predict the Crime Risky Areas without crime data. As the first step, we analyzed the three factors affecting crime occurrence : demographic characteristics, building structure and housing types characteristics.
In this study, a probability statistic model was used to quantitatively assess the probability of crime occurrence and to predict the Crime Risky Area. In addition, this study used GIS to make it easier for citizens to identify Crime Risky Areas.
The most important purpose of this study is to identify the dangerous area of crime by citizens themselves. Therefore, this study use spatial big data provided by a public institutions free to citizens.
First, multiple regression analysis was used to select variables that affect crime occurrence among the spatial big data of crime safety. The Bayesian model was constructed using the significant variables to predict the Crime Risky Areas. The most dangerous areas in the predicted area were college campuses, and many crimes actually occurred. Also, the floating population is higher than the average.
In order to verify the predicted results, the study compared the actual crime spot and crime density analysis results. Also, the ROC(Receiver Operation Characteristic) curve was used to quantitatively verify the prediction accuracy, and the AUC(Area Under Curve) value showed a prediction accuracy of 86.1 percent.
Based on the above results, the crime occurrence risk was higher in area where residential and commercial facilities are concentrated and floating population is high. Therefore, in this area, measures such as revising and adding police patrol routes and installing street security bells are necessary. In addition, there is a need to identify the causes of frequent crime even though it represents high level of floating population and to take appropriate measures against it.
The results of this study can predict the Crime Risky Area by using the spatial big data without fear of leakage of personal information, and it is expected that the efficiency of citizen's own crime prevention will be able to be improved by utilizing this. In addition, since the characteristics of the Crime Risky Area are analyzed, it can be used as a guideline for preparing countermeasures for crime prevention and expected to contribute to building safer cities.
Society has been constantly changing due to the rapid growth of cities and the rise of economic levels. Along with this social change, crime which one of the threats to society, continues to increase. The crime has become more severe over time. As the crime diversified and the crime incidence increased, the need for a predictive crime prevention strategy attracted attention. In the case of foreign countries, most countries provide accurate information on crime data to predict Crime Risky Areas and minimize crime damage. However, the accurate information has not been provided to citizen in korea due to the leakage of personal information. Furthermore, it provides only crime occurrence area information of the global level. For this reason, Such information was less effective to korea citizen in predicting and preparing Crime Risky Areas. Therefore, in order to improve the efficiency of citizen's own crime prevention in Korea, it is necessary to prepare a plan to minimize crime damage without actual crime data. This study explored to predict the Crime Risky Areas without crime data. As the first step, we analyzed the three factors affecting crime occurrence : demographic characteristics, building structure and housing types characteristics.
In this study, a probability statistic model was used to quantitatively assess the probability of crime occurrence and to predict the Crime Risky Area. In addition, this study used GIS to make it easier for citizens to identify Crime Risky Areas.
The most important purpose of this study is to identify the dangerous area of crime by citizens themselves. Therefore, this study use spatial big data provided by a public institutions free to citizens.
First, multiple regression analysis was used to select variables that affect crime occurrence among the spatial big data of crime safety. The Bayesian model was constructed using the significant variables to predict the Crime Risky Areas. The most dangerous areas in the predicted area were college campuses, and many crimes actually occurred. Also, the floating population is higher than the average.
In order to verify the predicted results, the study compared the actual crime spot and crime density analysis results. Also, the ROC(Receiver Operation Characteristic) curve was used to quantitatively verify the prediction accuracy, and the AUC(Area Under Curve) value showed a prediction accuracy of 86.1 percent.
Based on the above results, the crime occurrence risk was higher in area where residential and commercial facilities are concentrated and floating population is high. Therefore, in this area, measures such as revising and adding police patrol routes and installing street security bells are necessary. In addition, there is a need to identify the causes of frequent crime even though it represents high level of floating population and to take appropriate measures against it.
The results of this study can predict the Crime Risky Area by using the spatial big data without fear of leakage of personal information, and it is expected that the efficiency of citizen's own crime prevention will be able to be improved by utilizing this. In addition, since the characteristics of the Crime Risky Area are analyzed, it can be used as a guideline for preparing countermeasures for crime prevention and expected to contribute to building safer cities.
주제어
#범죄 예측 공간 빅데이터 베이지안
※ AI-Helper는 부적절한 답변을 할 수 있습니다.