短时交通流预测方法综述

Summary of Short-time Traffic Flow Forecasting Methods

以交通流预测研究的步骤为主线,对短时交通流预测的方法进行研究。对现存预测方法进行了分类分析:基于统计理论的方法、基于神经网络的方法、基于非线性理论的方法以及基于新兴技术的预测方法。将人工神经网络模型与其他领域的研究相结合的综合预测模型要比单一神经网络预测模型、常规预测模型的预测效果好;基于非线性理论的预测方法有较好的发展前景。

This article follows the steps of the intelligent transport research as its main stream, and gives us a summary of the research methods on short -term traffic flow forecasting. It analyzes the existing forecasting methods from the following four aspects：methods based on statistical theory, methods based on neural network, methods based on non - linear theory, and forecasting methods based on the rising technology. The integrated forecasting model which joins the artificial neural network to researches in other areas has a better forecasting effect than the single model of neural network and the regular forecasting model. Also, the forecasting method which based on non - linear theory will have a better developing prospect.