【摘要】 为了准确预测天津快速路的交通运行速度，提出基于变分模态的长短期(LSTM)神经网络预测模型.首先，采用变分模态算法，对原始交通速度数据进行分解，将原始数据分解成多个子模态以降低原始数据波动性对预测精度产生的影响.其次，依据不同运行速度子模态建立LSTM神经网络，对后数小时的交通运行速度进行预测.最终，将各子模态的LSTM神经网络预测结果进行重构，得到快速路预测速度.预测结果证实，基于变分模态的LSTM预测模型预测精度优于其他预测模型，性能表现最优.更多还原

【Abstract】 In order to accurately predict the traffic speed of Tianjin expressways, a long short-term memory(LSTM) neural network prediction model based on variational mode decomposition is proposed. First, the original traffic speed data is decomposed into multiple sub-modes by using the variational mode decomposition to reduce the influence of the original data fluctuation on the prediction accuracy. Second, the LSTM models for predicting the traffic speed of next few hours are established. Finally, the prediction results of each sub-mode are reconstructed to obtain the speed of Tianjin expressways. The prediction results show that the prediction accuracy of LSTM model based on variational mode decomposition is better than other prediction models, and the LSTM prediction model performs the best in traffic speed prediction.更多还原