基于ARMA模型的我国财产类犯罪人数预测研究

发布时间：2018-11-07 15:45

【摘要】：在定量犯罪学的语境下,以《中国法律年鉴》全国检察机关决定批捕和自行决定批捕的侵犯财产类犯罪人数之和作为我国财产类犯罪人数的统计衡量指标,做出了1986-2010年全国财产类犯罪人数变迁的时间序列。通过时间序列分析法和计算机软件EViews5.0,对时间序列逐步进行平稳性检验、独立性检验,建立了我国财产类犯罪的ARMA(1,10)拟合预测模型。一方面,预测分析表明2011年我国财产类犯罪人数的点预测结果为387914,区间预测结果为[334369,451351](单位:人);另一方面,拟合精度分析证实,建立的模型通过了统计学检验,模型整体显著,预测精度较高,是一个优良的时间序列预测模型,对我国财产类犯罪定量实证预测研究能起到良好的借鉴作用。
[Abstract]:In the context of quantitative criminology, the sum of the number of crimes against property decided by the procuratorial organs of China and the number of crimes against property determined by the procuratorial organs of the whole country is taken as the statistical measure index of the number of property crimes in China. The time series of the change of the number of property crimes in China from 1986 to 2010 was made. By means of time series analysis method and computer software EViews5.0, the ARMA (110) fitting prediction model of property crimes in China is established by means of time series analysis method and computer software EViews5.0, to test the stability and independence of the time series step by step. On the one hand, the forecast analysis shows that the point forecast result of property crime in 2011 is 387914, the interval forecast result is [334369 / 451351] (unit: person); On the other hand, the analysis of fitting precision proves that the established model has passed the statistical test, and the model is remarkable as a whole and has high prediction accuracy, so it is an excellent time series prediction model. It can be used for reference in quantitative empirical prediction of property crimes in China.
【作者单位】：湖南大学法学院;
【基金】：2012年湖南大学国家级SIT项目“基于数据年鉴的我国财产类犯罪变迁研究”(531107061080)
【分类号】：D917

【参考文献】