基于AGA-PP的农村公共安全综合风险评价方法

General Risk Evaluation Method of Rural Public Security based on AGA-PP

为了构建一套有效评价农村公共安全风险的指标体系和评价方法,考虑农村环境承载力影响因素对农村公共安全的制约,围绕农村生态承载力、基础设施承载力、社会承载力和人口承载力4个维度,提出能够全面评价农村公共安全风险的指标体系和基于PP(projection pursuit,投影寻踪)模型的风险评价方法.在风险评价方法构建上,采用AGA(accelerated genetic algorithm,加速遗传算法)对PP模型进行改进,实现投影指标函数的快速优化,在此基础上,采用回归分析方法,将时间序列与一维投影值进行拟合,求得一维投影值的预测曲线,通过对一维投影值的预测,利用信息扩散理论将投影值所携带的风险信息扩散至设定的4个风险等级上,用风险熵计算农村环境承载力发展趋势,并借以评价农村公共安全的风险大小.结果表明:基于AGA-PP的评价方法能够最大限度暴露多维指标数据的结构特征.依据最佳投影方向获得的一维投影值与时间序列拟合后,其拟合精度达到0.027 1,较高的拟合精度保证了后续农村环境承载力风险熵值的可信度,这在一定程度上保证了农村公共安全风险评价的可靠性.研究显示,从农村环境承载力的视角评价农村公共安全风险能够很好地反映农村公共安全风险的现实状况和发展趋势.基于AGA-PP的农村公共安全评价方法简单有效,具有评价精度高且与实际吻合度高的特征,能够为公共安全管理部门提供决策依据和方法参考.

To construct a set of effective evaluation index system and evaluation method of rural public security risk,which is the basic work of promoting rural public security construction scientifically.The paper considers the rural environment influence factors on the bearing capacity of rural public safety around 4 dimensions of bearing capacity of rural ecological,infrastructure,social and population,and puts forward the evaluation indexes and the projection pursuit(PP)risk assessment model of the rural public safety risk.The paper uses the accelerated genetic algorithm(AGA)to improve the standard projection pursuit method to achieve rapid optimization of projection index function,on this basis,the paper also uses the regression analysis method to fit between time series and one-dimensional projection value to get the prediction curve of one-dimensional projection value,and then the projection values carried by the theory of information diffusion to spread into the 4 risk levels.This paper uses the method of risk entropy to calculate the development trend of rural environmental carrying capacity for evaluating the risk size of rural public security.The verification results show that the improved PP evaluation model can expose the structural characteristics of multidimensional index to the maximum extent.The author fits between one-dimensional projection value and time series,and find the fitting precision reaching 0.0271,and therefore,high fitting precision can ensure the rural environmental bearing capacity risk entropy credibility,so that also ensures the assessment result of the rural public safety risk assessment reliably in a certain extent.The conclusion is:that the rural public security risk assessment can reflect the actual situation and development trend of rural public safety risk around the perspective of rural environmental carrying capacity.The assessment model is a simple and effective method with high precision,stable effect and consistent with reality,so it can provide decision basis and method reference for the department of public security management.