Research Article

Traffic Flow Prediction and Analysis in Smart Cities Based on the WND-LSTM Model

Table 2

Comparison of evaluation metrics between WND-LSTM and other advanced models.

DaysMOEsModels
ARIMALRSVRKNNSAEsGRULSTMWND-LSTM

1 dayMAPE5.2972.5583.4932.4584.4734.6346.1222.159
MAE0.0950.040.0620.0470.0780.0760.1370.056
RMSE0.1230.0050.080.0660.0840.0920.1840.096
ME0.3020.1160.1630.2450.1540.2190.3690.372

3 daysMAPE6.0312.8196.7014.2972.3084.6562.5871.795
MAE0.1380.0670.2730.160.0490.1310.1450.053
RMSE0.1960.0960.7640.40.0690.250.2750.102
ME0.6890.4383.11.6290.2690.9721.7220.410

6 daysMAPE3.8351.9462.4662.0372.2051.9882.7560.75
MAE0.0560.0300.0340.0310.0320.0320.0440.012
RMSE0.0720.0360.0410.0380.0380.0420.0590.015
ME0.1820.0860.0950.0940.0870.1140.1640.039

10 daysMAPE5.8421.9522.5092.0612.5152.0162.9880.716
MAE0.1120.0380.0420.040.0520.0390.0450.015
RMSE0.1390.0470.050.050.0660.0520.0540.018
ME0.420.1440.1460.1670.1550.1660.1410.048

14 daysMAPE5.6521.9032.2042.0461.972.0652.20.51
MAE0.0840.0280.030.030.0290.0310.030.01
RMSE0.1060.0360.0380.040.0370.0410.0370.014
ME0.2980.10.1130.120.1030.1060.1050.044