论文部分内容阅读
根据交通流量具有周相似的特性,利用实地采集获取的交通流量数据,选取每周周一至周五的数据,构造时间序列。本文分别用了3种不同的方法(BP神经网络、Elman神经网络、RBF神经网络)来预测短时交通流量,并通过不同的评价指标对上述3种方法的预测效果进行了评价。实例分析表明,对于这种时间序列的预测,Elman神经网络预测效果优于其他2种方法,更适合于短时交通流预测。
According to the similar characteristics of traffic flow in the week, the traffic flow data collected in the field are used to select the data from Monday to Friday of each week to construct the time series. In this paper, three different methods (BP neural network, Elman neural network, RBF neural network) were used to predict short-term traffic flow, and the prediction results of the above three methods were evaluated by different evaluation indexes. The case study shows that Elman neural network is better than the other two methods in predicting this kind of time series and is more suitable for short-term traffic flow prediction.