[1]李香云,任 帅,张卫钢,等.基于高斯过程回归的公交到站预测方法[J].计算机技术与发展,2019,29(10):21-25.[doi:10. 3969 / j. issn. 1673-629X. 2019. 10. 005]
　LI Xiang-yun,REN Shuai,ZHANG Wei-gang,et al.A Bus-to-station Prediction Method Based on Gaussian Process Regression[J].,2019,29(10):21-25.[doi:10. 3969 / j. issn. 1673-629X. 2019. 10. 005]
点击复制

基于高斯过程回归的公交到站预测方法()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
29
期数:
2019年10期
页码:
21-25
栏目:
应用开发研究
出版日期:
2019-10-10

文章信息/Info

Title:
A Bus-to-station Prediction Method Based on Gaussian Process Regression
文章编号:
1673-629X(2019)10-0021-05
作者:
李香云; 任 帅; 张卫钢; 吴娟娟; 伍 菁
长安大学 信息工程学院,陕西 西安 710064
Author(s):
LI Xiang-yun; REN Shuai; ZHANG Wei-gang; WU Juan-juan; WU Jing
School of Information Engineering,Chang’an University,Xi’an 710064,China
关键词:
城市交通; 高斯过程回归; 精准预测; 置信区间; 概率性预测
Keywords:
urban traffic; Gaussian process regression; precise prediction; confidence interval; probabilistic prediction
分类号:
TP301
DOI:
10. 3969 / j. issn. 1673-629X. 2019. 10. 005
摘要:
在城市交通管理中,向市民提供更多的公共交通服务,是实现节能减排的有效途径。 而研究公交车到站预测问题对提高公共交通服务水平有着重要的现实意义。 目前,在公交车到站时间预测的研究中,大都是围绕到站时间精准性问题,而没有对预测结果的不确定性进行定量分析。 因此,文中提出了一种基于高斯过程回归的公交到站预测方法,不仅可以对公交车到站时间进行精准预测,还可以根据预测值的方差来确定预测值 95% 的置信区间,即对不确定性进行研究。实验结果表明,该预测方法不仅与基于支持向量机的预测方法具有相近的预测精度,其中标准误差为 13.39,平均绝对误差为 12.91,平均绝对百分比误差为 0.14,而且能够有效实现公交车到站时间概率意义上的预测。
Abstract:
In urban traffic management,providing more public transportation services to the public is an effective way to achieve energy conservation and emission reduction. The research on bus arrival prediction is of important practical significance for improving the public transportation service level. At present,most of the research on the arrival time prediction of buses is based on the accuracy of the arrival time,but there is no quantitative analysis of the uncertainty of the prediction results. Therefore,we propose a bus-to-station prediction method based on Gaussian process regression,which can not only accurately predict the arrival time of the bus,but also determine the 95% confidence interval of the predicted value based on the variance of the predicted value,that is,the uncertainty is studied. The experiment shows that the prediction method proposed not only has similar prediction accuracy with the prediction method based on support vector machine,in which the root mean square error is 13.39,the mean absolute error is 12.91,and the mean absolute percentage error is 0.14,but also can effectively realize the prediction of the probability of bus arrival time.

相似文献/References:

[1]张宝磊 任军号 巩岁平.出行分布预测模型及其系数标定算法研究[J].计算机技术与发展,2011,(04):21.
　ZHANG Bao-lei,REN Jun-hao,GONG Sui-ping.Forecast Model of Trip Distribution and Coefficient Calibration Algorithm[J].,2011,(10):21.
[2]傅赟,王桂丽,周旭廷,等. 交通监控系统中视频运动目标检测算法研究[J].计算机技术与发展,2017,27(08):156.
　FU Yun,WANG Gui-li,ZHOU Xu-ting,et al. Investigation on Video Moving Target Detection Algorithm in Traffic Monitoring System[J].,2017,27(10):156.
[3]梅朵[],郑黎黎[],刘春晓[],等. 基于混合算法优化SVM的短时交通流预测[J].计算机技术与发展,2017,27(11):92.
　MEI Duo[],ZHENG Li-li[],LIU Chun-xiao[],et al. A Short-term Traffic Flow Prediction Model Based on Support Vector Machine Optimized by Hybrid Algorithm[J].,2017,27(10):92.
[4]刘 锂,段芃芃.智慧城市交通网络系统的研究与应用[J].计算机技术与发展,2019,29(08):185.[doi:10. 3969 / j. issn. 1673-629X. 2019. 08. 035]
　LIU Li,DUAN Peng-peng.Research and Application of Smart City Traffic Network System[J].,2019,29(10):185.[doi:10. 3969 / j. issn. 1673-629X. 2019. 08. 035]

更新日期/Last Update: 2019-10-10