基于完全自适应噪声集合经验模态分解的短时交通流组合预测

Short-term Traffic Flow Combined Prediction Based on Complementary Ensemble Empirical Mode Decomposition with Adaptive Noise

为了提高短时交通流预测的准确性,鉴于短时交通流非平稳、难预测的特征,提出了基于完全自适应噪声集合经验模态分解(CEEMDAN)短时交通流组合预测方法。利用CEEMDAN将原始短时交通流信号进行分解得多个复杂度、频率不同的时间序列分量,利用排列熵算法(PE算法)计算各分量的复杂度;然后根据复杂度和随机性的不同分为高频和低频,分别使用ATT-TCN-BIGRU模型和ARIMA模型对高频分量和低频分量进行预测,最后叠加高频和低频的每个分量预测结果作为最终短时交通流预测值。仿真分析结果表明:与ARIMA模型、TCN模型、BIGRU模型、ATT-TCN-BIGRU模型相比,此模型的平均绝对误差及平均绝对百分比误差为最小,预测精度更高。

In order to improve the accuracy of short-term traffic flow prediction,a short-term traffic flow combination prediction method based on Complementary Ensemble Empirical Mode Decomposition with Adaptive Noise(CEEMDAN)is proposed in view of the non-stationary and unpredictable characteristics of short-term traffic flow.CEEMDAN is used to decompose the original short-time traffic flow signal into multiple time series components with different complexity and frequency,and the complexity of each component is calculated by permutation entropy algorithm(PE algorithm).Then,according to the complexity and randomness,it is divided into high frequency and low frequency,and the ATT-TCN-BIGRU model and ARIMA model are used to predict the high frequency component and low frequency component,and finally the prediction results of each component of high frequency and low frequency are superimposed as the final short-term traffic flow prediction value.The simulation results show that compared with the ARIMA model,TCN model,BIGRU model and ATT-TCN-BIGRU model,the average absolute error and average absolute percentage error of this model are the smallest,and the prediction accuracy is higher.