Abstract: In the face of the increasingly diverse needs of people travel, more and more customized public transport appears in the peoples life. Before planning a new type of public transport which means to provide the customized travel services for target passengers, it is necessary to predict the target passengers travel intentions. Considering the complexity of the influence factors of travel intention, a predict method based on wavelet neural network was proposed. First of all, according to the survey of passengers travel intention, the factors affecting the target passengers travel intention were analyzed; and then the neural network model was constructed.Finally, the forecast was verified and implemented by using MATLAB. In order to improve the training effect of the network, the self-adaptive method was used to determine the networks learning rate and the number of hidden layer nodes.The results show that the wavelet neural network has good applicability for the prediction of target passengers travel intention of the customized public transport, and has better prediction accuracy than the traditional BP neural network.

Key words: traffic engineering, customized public transport;travel intention, wavelet neural network, BP neural network

摘要： 面对人们日益多样化的出行需求，越来越多的定制化公共交通出现在了人们生活中。在规划一种意为目标乘客提供定制化出行服务的新型公共交通方式之前，需要对目标乘客对其的出行意愿进行准确预测。基于出行意愿的影响因素的复杂性，本文提出了一种基于小波神经网络的预测方法。首先，根据目标乘客出行意愿调查进行出行意愿的影响因素分析，然后对神经网络模型进行构建，最后运用MATLAB对预测进行了实现、验证。为了改善网络的训练效果，本文采用自适应的方法对网络学习速率和隐含层节点数进行确定。结果表明，小波神经网络对于预测定制公交目标乘客出行意愿具有较好的适用性，且比传统的BP神经网络有着更好的预测精度。

关键词: 交通工程, 定制公交, 出行意愿, 小波神经网络, BP神经网络