Table 2 Performance comparison of deep learning models (LSTM, GRU, and RNN) over different Epochs.
Model | Epochs # | Train Acc | Valid Acc | Train Loss | Valid Loss | Time (sec) |
|---|---|---|---|---|---|---|
LSTM | 1 | 51.52 | 40.28 | 1.35 | 1.50 | Â |
5 | 84.50 | 80.28 | 0.50 | 0.65 | 7223 | |
10 | 91.20 | 89.28 | 0.23 | 0.43 | Â | |
15 | 91.20 | 89.60 | 0.21 | 0.40 | Â | |
GRU | 1 | 49.52 | 45.28 | 1.40 | 1.55 | Â |
5 | 84.00 | 81.28 | 0.55 | 0.70 | Â | |
10 | 90.00 | 88.80 | 0.34 | 0.50 | 2281 | |
15 | 90.25 | 89.14 | 0.30 | 0.45 | Â | |
RNN | 1 | 45.28 | 40.25 | 1.45 | 1.60 | Â |
5 | 80.50 | 75.50 | 0.60 | 0.75 | 4225 | |
10 | 88.90 | 86.50 | 0.39 | 0.54 | Â | |
15 | 89.01 | 87.39 | 0.35 | 0.49 | Â |
