摘要
道路交通伤害发生率高,造成伤害严重,已成为全球性公共卫生问题.道路交通伤害预测对掌握其未来的发生状况,及时采取相应的措施,具有重要作用.目前常用的预测方法主要有回归分析法、时间序列法、灰色模型法和BP(Back Propagation,BP)神经网络法等.回归分析法主要适用于样本量大、数据波动小、规律性较强的预测;时间序列分析法中应用较多的是自回归滑动平均混合模型(autoregressive integrated moving average model,ARIMA),其对数据的要求不高,具有较好的拟合效果,适用于近期或短期预测;灰色模型法所需样本数据少、原理简单、运算方便,特别适合对具有复杂性、随机性和灰色性特点的道路交通伤害的预测;BP神经网络法则需要较为全面的数据,短期预测精度高,可用于宏观问题的预测.本文就道路交通伤害预测方法进行综述,总结各方法的适用条件及优缺点,为减少和预防道路交通伤害提供依据.
The road traffic injuries occru at high incidences and cause serous consequences thua have become a seriouspublic health problem worldwide ,The existing,commonly used predicting methods include regression analysis.time series method,grey model method. and BP(Back Propagation ,BP)neural network,Regression analysis is suitable for the prediction of large samles with small fluctuations and a profound regularity.ARIMA(autoregressive integrated moving average model, ARIMA)model is widely applied in time series analysis with acceptable fitting effects.Comprehensive understanding and careful application of these methods may reduce the incidence of road traffic injuries.
出处
《伤害医学（电子版）》
2013年第2期49-53,共5页
Injury Medicine(Electronic Edition)
关键词
道路交通伤害
预测方法
文献综述
Road traffic injuries
Predicting methods
Literature reviewarima
