融合多变量主成份与因子分析算法的犯罪率研究

Research on Crime Rate with Integration of Multivariate Principal Component and Factor Analysis Algorithm

基于犯罪过程的不确定性结构,融合多变量主成份分析与因子分析算法,分析了影响犯罪率的城市化率、失业率、客运量、人均GDP、城市基尼系数、农村基尼系数等多个相关变量因素,由MATLAB仿真建立了犯罪率的多元回归模型,并采用MINITAB数学软件对浙江省1988-2007二十年样本数据进行了实测,获取了影响犯罪率的潜在因子与方差贡献率,讨论了多元回归模型的统计检验及相对误差,解决了犯罪率各相关变量的权重分布及信息相互重叠问题.

Based on the uncertainty structure of criminal process, integration of multivariate principal component and factor analysis algorithm has analyzed correlative variables which influence on crime rate such as urbanization rate, unemployment rate, passenger capacity, GDP per capita, urban GINI coefficient, country GINI coefficient, and established multiple regression model to forecast crime rate based on MATLAB simulation platform, measured the sample data in two decades from 1988 to 2007 of Zhejiang province with adoption of MINITAB mathematical software, obtained potential factors affecting crime rate and variance devoting rates, discussed statistical test or relative error of multiple regression model of forecast crime rate, and solved weight distribution or overlapped information of correlative variables on crime rate.