摘要
为分析道路交通状态波动范围,提出了一种基于模糊信息粒化与支持向量机组合预测的建模方法。分析了道路交通状态波动特点和交通参数选择原则,以模糊理论和时间序列预测为基础,通过模糊信息粒以15 min时间窗将样本数据模糊化,得到Low、R、Up这3组时间序列;并利用支持向量机模型分别对其进行预测,获得道路交通状态的波动范围与变化趋势。实例分析时,在验证数据采集路段属性相近的前提下,用该组合模型对早、晚高峰和平峰等3个时段的交通波动状态进行验证,验证结果有较高精度,能有效预测交通状态波动情况。
In order to analyze the range of traffic state fluctuation, a combined modeling method of forecasting the range of traffic state fluctuation based on fuzzy information granulation and support vector machine was put forward. The fluctuation characteristics of road traffic state and the selection principles of traffic parameters were analyzed on the basis of fuzzy theory and time series prediction. The sample data were fuzzed by 15 min time window through the fuzzy information granules, and three sets of time series, including Low, R and Up, were obtained. The support vector machine was used to forecast each set of time series, and the traffic state fluctuation range and variation tendency were obtained. In case studies, in the premise of verification data acquisition segment with similar attributes, the proposed model was used to analyze the fluctuation of traffic state in morning, evening peak periods and peace period, whose results were of high accuracy and could effectively predict the fluctuation of traffic state.
出处
《重庆交通大学学报（自然科学版）》
CAS
北大核心
2017年第7期83-89,95,共8页
Journal of Chongqing Jiaotong University(Natural Science)
关键词
交通工程
交通状态
波动范围
模糊信息粒化
支持向量机
traffic engineering
traffic state
fluctuation range
fuzzy information granulation
support vector machine
