摘要
为提高公交客流量预测的精确度,将混沌理论和小波神经网络方法相结合应用于公交客流量预测。分别采用自相关法、伪最近邻域法计算公交客流量时间序列的时间延迟、嵌入维数,采用小数据量法计算其最大李雅普诺夫指数,证实该时间序列具有混沌特性。据此建立混沌-小波神经网络预测模型,进而对H省某市实际公交客流量进行预测。实验结果表明,相比于传统的BP神经网络预测法、LIBSVM预测法,该方法在均方误差(MSE)、平均绝对误差(MAE)、平均相对误差(MRE)上均具有更小的预测误差,因而可以有效地预测公交客流量。
In order to improve the accuracy of public transportation passenger volume prediction, this article integrates chaos theory and wavelet neural network method into the prediction. We deploy autocorrelation method and false nearest neighbor domain method to calculate the delay time and embedding dimension of the public transportation passenger volume time series. After that, by utilizing small data sets method, we obtain its largest Lyapunov exponent and then the chaotic characteristic of the time series is proved. According to this, we establish the chaos-wavelet neural network prediction model to predict the actual public transportation passenger volume. Finally, by applying our method on the real data set from some city of H Province, the experimental results demonstrate that our approach achieves the smaller prediction error on Mean Square Error （MSE）, Mean Absolute Error （MAE） and Mean Relative Error （MRE） compared with the traditional prediction methods, such as BP neural network and LIBSVM. So it is able to predict public transportation passenger volume effectively.
出处
《城市公共交通》
2017年第9期34-40,共7页
Urban Public Transport
关键词
混沌
小波神经网络
公交客流量预测
chaos
wavelet neural network
public transportation passenger volume prediction
