摘要
实时、准确的短时交通流预测是交通控制与诱导中的一个关键问题和难点.非参数回归是解决短时交通流预测问题的较好方法,但是案例库生成难和搜索速度慢是其目前实际应用的两大障碍.为此,提出一种基于平衡二叉树的K-邻域非参数回归(KNN-NPR)的短时交通流预测方法,采用聚类方法和平衡二叉树结构建立案例数据库,以提高预测精度和满足实时性要求.给出了预测示例说明了方法的有效性.
Real-time and accurate short-term traffic flow forecasting has become critical in traffic control and guidance. Non-parametric regression is a good way to solve the problem. But the foundation of case database and the search speed are two obstacles for application. A KNN-NPR（ K-nearest neighbors non-parametric regression） method based on balanced binary tree to forecast short-term traffic flow is presented. In order to improve forecasting precision and meet real-time reqirement, clustering methods and balanced binary tree are adopted to build case database. An example is given to show its availability.
出处
《系统工程学报》
CSCD
北大核心
2009年第2期178-183,共6页
Journal of Systems Engineering
基金
北京市科学技术委员会科技计划资助项目(D07020601400705)
关键词
短时交通流预测
非参数回归
聚类
平衡二叉树
short-term traffic flow forecasting
NPR （ non-parametric regression）
clustering
balanced binary tree
