Table 9.

Baseline and State-of-the-Art Comparison.

Model	Accuracy (%)	F1-Score	MAE	Inference Time (ms)	Memory (GB)
Proposed (TCN + BiLSTM + Attention Mechanism)	93.45	0.924	0.187	12.34	3.86
LSTM	87.23	0.862	0.284	15.67	2.94
GRU	86.58	0.854	0.296	14.82	2.87
1D-CNN	85.92	0.847	0.312	11.45	2.56
DeepConvLSTM	89.76	0.886	0.245	16.93	3.42
Transformer	91.23	0.903	0.214	18.56	4.12