论文部分内容阅读
实时准确的短时交通流预测在城市道路交通和高速公路交通中都十分重要,是交通控制与诱导系统的基础。应用在线支持向量回归算法对交通流进行预测,并对济南某高架路实测数据进行仿真运算。预测结果表明,在小样本下,与BP神经网络算法相比,在线支持向量回归算法明显优于BP神经网络算法,增大样本数,BP神经网络算法预测精度有所提高,但仍低于线支持向量回归算法;在运算时间上,BP神经网络算法运算时间更短。
The real-time and accurate short-term traffic flow prediction is very important in urban road traffic and expressway traffic, which is the basis of traffic control and guidance system. Apply online support vector regression algorithm to forecast traffic flow and simulate the measured data of an elevated road in Jinan. The prediction results show that the proposed algorithm is superior to BP neural network algorithm in comparison with BP neural network algorithm under small sample size. The prediction accuracy of BP neural network algorithm is improved, but it is still lower than that of BP neural network algorithm Support vector regression algorithm; BP neural network algorithm in computing time, computing time is shorter.