Table 4. Threshold analysis of dynamic supervised classifier.
Model+Abstract Layers | Number of vectors | RBM1(%) | RBM2(%) | RBM3(%) | RBM4(%) | Average(%) | Standard deviation(%) |
---|---|---|---|---|---|---|---|
DBN4+ POSL+DPL+ LENL+LOCL+ DISL+TWFL | |||||||
2000 | 7.25 | 4.17 | 2.61 | 2.52 | 4.14 | 2.21 | |
4000 | 7.18 | 4.09 | 2.52 | 2.49 | 4.07 | 2.2 | |
6000 | 7.13 | 4.09 | 2.51 | 2.48 | 4.05 | 2.19 | |
8000 | 7.12 | 4.11 | 2.52 | 2.45 | 4.05 | 2.19 | |
10000 | 7.08 | 4.01 | 2.46 | 2.42 | 3.99 | 2.19 | |
12000 | 7.01 | 3.99 | 2.45 | 2.46 | 3.98 | 2.15 | |
14000 | 6.98 | 3.97 | 2.42 | 2.41 | 3.95 | 2.15 | |
16000 | 6.97 | 3.93 | 2.38 | 2.41 | 3.92 | 2.16 | |
18000 | 6.9 | 3.89 | 2.35 | 2.36 | 3.88 | 2.14 | |
20000 | 6.86 | 3.91 | 2.38 | 2.32 | 3.87 | 2.13 | |
Average value | 11000 | 7.048 | 4.016 | 2.46 | 2.432 | 3.99 | 2.17 |