论文部分内容阅读
针对公路通道交通量预测的复杂性,提出偏最小二乘神经网络组合模型。传统的预测方法需要考虑城市经济、人口、工业发展等因素,这些因素之间存在密切的相关性,往往使得预测精度降低。本文将针对这些不足进行改进,将偏最小二乘方法和改进神经网络方法相结合,提取对因变量解释最强的成分进行预测。将该预测模型应用于揭阳市公路网规划通道交通量预测中,取得了良好的效果。
Aiming at the complexity of highway traffic volume forecasting, a partial least squares neural network combined model is proposed. The traditional forecasting methods need to consider such factors as the urban economy, population and industrial development. These factors are closely related and tend to reduce the prediction accuracy. In this paper, we will improve these shortcomings, combine the partial least squares method with the improved neural network method, and extract the components with the best explanation of the dependent variable. Applying the forecasting model to the forecast of the traffic volume of the highway network planning channel in Jieyang City has achieved good results.