基于高斯核函数的短时交通流量预测

Short-term Traffic Flow Forecast Based on the Gaussian Kernel

交通出行在人们的生活中占据了重要的地位,准确、短时交通流量预测有助于实现交通控制.本文采用基于高斯核函数的Nadaraya-Watson估计方法对短时交通流量进行了预测.以所选取路段的车流高峰时段为研究对象,根据当前时刻的车流量对下一时刻的流量值进行预测,其中相邻两时刻的时间间隔为5分钟;就所选的高斯核函数而言,根据试凑法,选定核函数带宽为8.结果表明,该方法的预测效果良好,其均等系数为0.996545243,达到了很好的预测效果.

Traffic plays an important role in our live.Short-term traffic flow forecast to contribute to the achievement of the traffic control.This article use the Nadaraya-Watson estimation method based on Gaussian Kernel function to predict short-time traffic flow.We use the traffic volume of the current moment to predict the volume of next moment.By the way,the interval is 5 minutes.Based on trial and error method,we selected kernel function bandwidth of 8.As we can see,the equalization coefficients of it is 0.996545243,this method has good prediction effect.