摘要
提出了一种基于相关向量机建模与求解思想的短时交通时序数据平滑处理方法,对短时交通时序参数平滑处理方法的流程进行了设计,并选取均方根误差、模型训练时间等作为评价指标。以西安市南二环快速路在不同时间尺度下的短时交通量实测数据对该平滑处理方法的有效性进行了验证,结果表明,提出的数据平滑处理方法可为分析城市短时交通时序参数态势变化规律、提高其参数预测精度等方面提供科学方法和理论支撑。
A smoothing method of short-term traffic timing parameters based on relevance vector machine modeling and solution is proposed. The step of smoothing method of short-term traffic timing parameters is designed,and root-mean-square error as well as model training time are selected as evaluating indicator. The effectiveness of the smoothing method is confirmed by the traffic volume measured data at different time scales in Middle of South 2nd-ring Road. Results indicate that the smoothing method provides scientific methods and theoretical support for analyzing the trend of variation of short-term traffic timing parameters as well as improving its forecast accuracy.
出处
《西北大学学报(自然科学版)》
CAS
CSCD
北大核心
2017年第1期38-42,共5页
Journal of Northwest University(Natural Science Edition)
基金
陕西省自然科学基础研究计划基金资助项目(2016JM5036)
陕西省交通科技基金资助项目(15-42R
15-39R)
关键词
交通工程
短时交通参数
时间序列
相关向量机
参数平滑
traffic engineering
short-term traffic parameters
time series
relevance vector machine(RVM)
smoothing parameter