BISINDO LSTM Model
Deskripsi
Model LSTM untuk mendeteksi Bahasa Isyarat BISINDO.
Hasil Training
- Final Training Loss: 6.4985
- Final Training Accuracy: 0.9586
| Epoch | Loss | Accuracy | Precision | Recall | F1-Score |
| ----- | ------- | -------- | --------- | ------ | -------- |
| 1 | 75.3358 | 0.0317 | 0.0041 | 0.0305 | 0.0068 |
| 2 | 74.9901 | 0.0400 | 0.0015 | 0.0385 | 0.0030 |
| 3 | 74.8302 | 0.0455 | 0.0034 | 0.0419 | 0.0063 |
| 4 | 74.7396 | 0.0676 | 0.0098 | 0.0625 | 0.0165 |
| 5 | 74.5590 | 0.0621 | 0.0296 | 0.0556 | 0.0168 |
| 6 | 74.1527 | 0.0897 | 0.0243 | 0.0811 | 0.0330 |
| 7 | 73.3582 | 0.0786 | 0.0159 | 0.0691 | 0.0215 |
| 8 | 71.8530 | 0.1297 | 0.1062 | 0.1164 | 0.0669 |
| 9 | 70.0197 | 0.1062 | 0.0479 | 0.0964 | 0.0478 |
| 10 | 67.8447 | 0.1241 | 0.1228 | 0.1113 | 0.0632 |
| 11 | 65.5784 | 0.1310 | 0.0843 | 0.1198 | 0.0743 |
| 12 | 63.5691 | 0.1614 | 0.1658 | 0.1486 | 0.0979 |
| 13 | 62.1602 | 0.1959 | 0.1285 | 0.1805 | 0.1213 |
| 14 | 60.0399 | 0.1986 | 0.2294 | 0.1857 | 0.1402 |
| 15 | 57.4834 | 0.2745 | 0.2789 | 0.2618 | 0.2071 |
| 16 | 55.2808 | 0.3131 | 0.2842 | 0.2995 | 0.2480 |
| 17 | 53.9088 | 0.3103 | 0.3817 | 0.3006 | 0.2729 |
| 18 | 51.7302 | 0.3559 | 0.3112 | 0.3427 | 0.3043 |
| 19 | 49.0403 | 0.3862 | 0.3847 | 0.3745 | 0.3267 |
| 20 | 46.0638 | 0.4483 | 0.4603 | 0.4382 | 0.4050 |
| 21 | 44.3573 | 0.4414 | 0.4632 | 0.4314 | 0.3902 |
| 22 | 42.4378 | 0.5214 | 0.5693 | 0.5091 | 0.4838 |
| 23 | 39.3630 | 0.5531 | 0.5799 | 0.5439 | 0.5202 |
| 24 | 38.4033 | 0.5600 | 0.5985 | 0.5501 | 0.5360 |
| 25 | 34.8533 | 0.6359 | 0.6514 | 0.6229 | 0.6057 |
| 26 | 33.1657 | 0.6510 | 0.6592 | 0.6384 | 0.6203 |
| 27 | 31.7321 | 0.6662 | 0.6850 | 0.6565 | 0.6490 |
| 28 | 29.5507 | 0.6883 | 0.6966 | 0.6817 | 0.6730 |
| 29 | 26.8903 | 0.7531 | 0.7630 | 0.7444 | 0.7387 |
| 30 | 25.7332 | 0.7366 | 0.7449 | 0.7291 | 0.7229 |
| 31 | 24.4123 | 0.7710 | 0.7718 | 0.7675 | 0.7638 |
| 32 | 23.0031 | 0.7600 | 0.7777 | 0.7563 | 0.7551 |
| 33 | 20.9617 | 0.8124 | 0.8246 | 0.8075 | 0.8068 |
| 34 | 19.3715 | 0.8262 | 0.8289 | 0.8233 | 0.8216 |
| 35 | 18.3695 | 0.8441 | 0.8450 | 0.8392 | 0.8390 |
| 36 | 15.9342 | 0.8786 | 0.8803 | 0.8769 | 0.8767 |
| 37 | 14.8239 | 0.8855 | 0.8882 | 0.8831 | 0.8841 |
| 38 | 14.3882 | 0.8745 | 0.8771 | 0.8732 | 0.8732 |
| 39 | 13.4870 | 0.8841 | 0.8873 | 0.8822 | 0.8829 |
| 40 | 12.0783 | 0.9062 | 0.9081 | 0.9041 | 0.9042 |
| 41 | 11.7186 | 0.9090 | 0.9114 | 0.9079 | 0.9080 |
| 42 | 11.1245 | 0.9172 | 0.9192 | 0.9159 | 0.9167 |
| 43 | 9.4605 | 0.9407 | 0.9416 | 0.9392 | 0.9388 |
| 44 | 9.1142 | 0.9352 | 0.9375 | 0.9350 | 0.9353 |
| 45 | 11.5496 | 0.8786 | 0.8787 | 0.8753 | 0.8758 |
| 46 | 11.3862 | 0.8800 | 0.8830 | 0.8792 | 0.8793 |
| 47 | 8.3005 | 0.9338 | 0.9353 | 0.9326 | 0.9334 |
| 48 | 6.6639 | 0.9683 | 0.9685 | 0.9673 | 0.9676 |
| 49 | 5.6791 | 0.9710 | 0.9721 | 0.9707 | 0.9711 |
| 50 | 6.4985 | 0.9586 | 0.9601 | 0.9577 | 0.9584 |
Hasil Evaluasi Test
Accuracy : 0.7143
Precision: 0.7144
Recall : 0.7075
F1-Score : 0.6987
ROC AUC : 0.9712
Visualisasi
