短时段交通预测时间序列方法的改进

Improvement of Time-Series Method for Short-Term Traffic Prediction

时间序列方法在短时交通预测中应用广泛.总结了该方法的优缺点.在传统的时间序列方法基础上引入带遗忘因子的最小二乘递推算法得到改进算法,实现了实时在线预测,解决了短时交通流量实时预测中存在的随机干扰因素影响大、不确定性强的问题.

The time-series method has come into wide use in short-term traffic prediction. This paper analys the merits and demerits of the method. On the basis of the traditional time-series method, an improved method is obtained by inducting the recursive forgetting factor least square method （RFFLS）, which brings about the real-time online prediction and smoothes away the problem of random severe interferences and great uncertainty in the real-time prediction of shortterm traffic flow.