Table 2.

Comparison of model performance under different configurations, highlighting the impact of applying smoothing filters.

Result	Model	RMSE	MSE	MAE		MAPE
Without filter	CNN	9.301	86.503	2.419	0.830	90.85
	LSTM	10.804	116.728	2.861	0.77	90.53
Kalman filter	CNN	9.669	93.484	2.504	0.816	90.64
	LSTM	11.943	142.63	3.308	0.719	90.09
Savitzky-Golay	CNN	8.789	77.253	2.262	0.848	92.22
	LSTM	11.409	130.159	3.159	0.744	91.08