A Hybrid Model for Short-Term Traffic Flow Prediction Based on Variational Mode Decomposition, Wavelet Threshold Denoising, and Long Short-Term Memory Neural Network

Table 4

The prediction errors of different models using the traffic flow data of detector B.

Prediction models	RMSE (veh/5 min)	MAE (veh/5 min)	MAPE (%)
LSTM	18.711	14.094	6.320
WD-LSTM	17.969	13.538	6.457
EMD-WD-LSTM	13.220	9.817	5.208
VMD-WD-ANN	49.357	37.466	13.365
EEMD-ANN	51.022	38.303	13.557
TSD-BiLSTM	18.765	13.664	6.531
ARIMA-LSTM	20.234	15.220	7.173
VMD-WD-LSTM	10.304	7.984	5.979

Bold indicates the best value of each evaluation index.