道路交通事故的三次指数平滑预测法.docx

文章编号：1008-0562(2014)01-0042-05doi:10.3969/j.issn.1008-0562.2014.01.009道路交通事故的三次指数平滑预测法王洪德，曹英浩（大连交通大学 土木与安全工程学院，辽宁 大连 116028）摘 要：为准确预测道路交通事故死亡人数和有效解决道路交通安全问题，以中国 2001~2012 年交通事故死亡 人数的统计数据为研究对象，首先，采用三次指数平滑法，建立三次指数交通事故预测模型，对道路交通事故数 据进行拟合与预测分析，绘制了中国道路交通事故的未来发展趋势图；然后，通过与一次指数平滑法、二次指数 平滑法进行对比分析，得出了三次指数平滑法预测精度较高，预测值与实际值相符合的结论；最后，预测出 2013年和 2014 年的道路交通事故死亡人数.研究结果表明：三次指数平滑法预测模型适用于短期预测，对有效预测道 路交通事故具有一定的参考价值和指导意义. 关键词：道路交通；死亡事故；预测模型；指数平滑法；初值选取；权重系数；时序非线性；对比分析 中图分类号：X 913.4文献标志码：ACubic exponential smooth method of road traffic accident forecastingWANG Hongde, CAO Yinghao（School of Civil safety Engineering, Dalian Jiaotong University, Dalian 116028, China） Abstract: In order to predict the death toll of road traffic accidents accurately and to solve the problem of road traffic safety effectively, the statistical data of death toll of road traffic accidents in China from 2001 to 2012 was considered as the research object. Firstly, the cubic exponential smoothing method was used to establish the traffic accidents forecasting model to fit and forecast the road traffic accident data, and to draw the developing trend of the road traffic accidents; Secondly, comparing with the single exponential smoothing method and the double exponential smoothing, the forecasting precision of the cubic exponential smoothing method are preferable, and the predicted values conformed to the true values; Finally, the forecasted death toll in road traffic accidents form 2013 to 2014 are 565065, 66883 respectively. Results show that the cubic exponential smoothing forecasting model is consistent with the characteristics of road traffic accidents, and can be used for short-term forecasting. The method can provide a certain reference and guidance for road traffic accident forecasting.Key words: road traffic; fatal accidents; forecasting model; exponential smoothing method; initial value selection; weight coefficient; temporal nonlinear; contrastive analysis0引言据统计，地球上平均每63秒就有1人死于