BP和RBF神经网络短时交通流预测结果比较  

Comparison of BP and RBF Neural Networks for Forecasting Short-term Traffic Flow

作　　者：胡洁[1] 

机构地区：[1]长江大学信息与数学学院,湖北荆州434023

出　　处：《长江大学学报（自科版）（上旬）》2007年第3期74-76,共3页JOURNAL OF YANGTZE UNIVERSITY (NATURAL SCIENCE EDITION) SCI ＆ ENG

摘　　要：对BP神经网络和RBF神经网络这2种模型的特征进行了分析,并将其应用于某高速公路的短时流量预测,比较了2种神经网络模型的预测结果。从量化的角度进一步证实了在交通流预测领域RBF神经网络比BP神经网络更快捷、更准确,从而更适合应用于对实时性和准确性要求比较高的交通系统。Based on artificial neural networks,the characteristics of BP,RBF neural networks are analyzed.The models and their features were compared and applied into short-term traffic volume prediction of a freeway,and the prediction results were compared.From a quantitative point of view,it is further confirmed that in traffic flow forecasting area RBF neural network is more efficient and more accurate than that of BP neural network,thus it is more suitable for real-time and high accuracy requirements of the transportation system.

关 键 词：短时交通流预测 BP神经网络 RBF神经网络 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]