Crime Prediction Research Based On Population Characteristics

With the rapid economic and social development, people's material life and spiritual civilization level has been significantly improved. In the early stage of economic development and social problems were not apparent, however, with the further development of economy, many social problems have become increasingly prominent. For example:the divorce rate and unemployment rising, imbalance between men and women, a widening income gap, population increase. At the same time, the number of criminal cases are increasing, so it is important to study the situation of criminal law, crime.At present, at home and abroad have on the relationship between economy and crime has been studied in many aspects, because the level of economic development directly influences the quality of life of residents quality of life, when the gap is too large, the social environment is relatively unstable, the lower the degree of social harmony, the contradiction is more easily excited, the quantity of crime is also increased. Due to China's large population, the collection of data is not convenient. Therefore the traditional study of crime often rely on experience or a single data to do qualitative analysis, hope that through the existing research results to take preventive measures, to reduce the incidence of criminal cases. But these methods often palliatives, on the one hand, usually study process requires a large amount of data, difficult collection; on the other hand, the long research cycle, resulting in research has been the rapid development of today does not apply. This pair of criminology research, especially the research on the crime situation, criminal law is very negative, but can not obtain complete and scientific conclusion. At the same time, the research on the traditional rarely has to analyze the problem from the quantitative point of view, can not fundamentally solve the problem, but also unable to crime prediction, this is also the significance of this study.The research contents of this paper, by using the grey correlation degree analysis and the advantage of qualitative research to pick out the characteristics of the population attributes and crime. On this basis, select the 7 attributes from more important, as a significant variable structural equation model, using AMOS 17.0, find out the quantitative relation between the number of population, the properties and the criminal cases and public security cases, before in this field, there are few quantitative research; in addition, this paper will weighted average combined with BP neural network is able to predict the future, the number of criminal cases.First of all, the author through the consult massive materials and sum up experience, there are a lot of, attributes that affect criminal cases such as:economic development level, educational level, employment problems, such as population density. In this paper, the number of criminal the case and the public security case as the dependent variable, to join the GDP per capita, the total population, the proportion of male, the floating population, the illiteracy rate, the rate of registered urban unemployment, the divorce rate, the Engel coefficient, the total dependency ratio and control variables as the employment population, carry out comprehensive statistical analysis, some of these properties will be bring a person pessimism, some will stimulate the circulation of money, may the birth of the phenomenon of crime in a certain extent.Secondly, the author determines the per capita GDP, population, illiteracy, the divorce rate, population and employment in the criminal cases and public security cases quantitative association degree is very high by grey correlation analysis and analysis of advantages, and the Engel coefficient of urban residents is caused by the increase in the number of criminal cases are the main factors to the total dependency ratio, and the number of cases of public security also has a great influence. Then, the author uses the structural equation model, and further to the attributes of the screened and two for the relationship between the variables of quantitative analysis.Finally, the author uses the attributes to predict the weighted average, and the 6 feature attributes as the input layer of the BP neural network, the criminal cases and public security cases respectively as the output layer of crime quantity forecast, the forecast precision is high, which shows that the model is reasonable and applicable, the promotion of the value.