为有效预测道路交通事故,促进交通可持续发展,利用支持向量机(SVM)与蚁群算法(ACA)综合研究道路交通事故问题。在对SVM进行优化决策的基础上,提出一个交通事故数据拟合和预测的统计分析模型。鉴于SVM的预测精度很大程度上取决于训练参数的选取,利用ACA优化其训练参数的选择过程,得到基于SVM的道路交通事故数据统计分析模型。利用该模型对小样本及非线性数据优越的预测性能进行年交通事故量的预测。结果表明,与一些其他模型相比,基于SVM的道路交通事故数据统计分析模型,预测精度更高、误差更小,能够更有效地对交通事故数量进行拟合、预测和统计分析。
In order to promote sustainable development of road traffic, road traffic accident problems were studied using support vector machines (SVM) and ant colony algorithm (ACA). A statistical analy- sis model for road traffic accidents was built. The model was verified by using data on road traffic accidents in China from 1995 to 2012. The results show that compared to other model can make much more accurate predictions.