摘要
为了提高出租车保有量的预测精度,利用小波神经网络逼近出租车保有量与其影响因素之间的非线性特性,并建立影响因素的预测模型,然后,将影响因素的预测值输入出租车保有量预测模型便实现了出租车保有量的预测。利用某市2000—2009年的出租车保有量及影响因素数据进行仿真预测,结果表明,相对于传统的BP神经网络,基于小波神经网络的出租车保有量预测模型具有更高的预测精度,该市2010—2012年的出租车保有量应分别达到9020、9 350、9 560 veh,才能保证平均候车时间在4 min左右。
To improve the prediction accuracy of the number of taxicabs, wavelet neural network was used to approximate the nonlinearity of the number of taxicabs with its influencing factors, and the prediction model of each influencing factor was established. Then, the predictive values of the influencing factors were input into the prediction model of taxicab number to predict the number of taxicabs in the future. The simulation was performed based on the data of the numbers of taxicabs and the influencing factors of a city from 2000 to 2009. The result shows that the model for predicting the number of taxicabs based on wavelet neural network has higher prediction accuracy than conventional BP neural network model, and the numbers of the taxicabs of this city from 2010 to 2012 should be 9 020, 9 350 and 9 560 respectively to keep the average waiting time around 4 min.
出处
《公路交通科技》
CAS
CSCD
北大核心
2012年第8期136-141,共6页
Journal of Highway and Transportation Research and Development
基金
国家自然科学基金项目(60978030)
关键词
交通工程
出租汽车保有量
小波神经网络
影响因素
预测模型
traffic engineering
number of taxicabs
wavelet neural network
influencing factor
prediction model
