Table 8 Comparison of the performance of the proposed method with other representations in the literature
Performance | ASCII | EIIP | One-hot | This work | ||
|---|---|---|---|---|---|---|
Metrics | [36] | [37] | Encoding [38] | \(M=64\) | \(M=128\) | \(M=256\) |
Processing time p/sequence | 0.0180 s | 0.0181 s | 0.648 s | 0.0063 s | 0.0063 s | 0.0064 s |
Memory required p/1000 vectors | 8.86 MB | 16.8 MB | 16.9 MB | 471 KB | 941 KB | 1.83 MB |
Training time per fold | 48.4 min | 46.55 min | 54.5 min | 12 s | 14 s | 17 s |
ACC | \(98\%\) | \(98.5\%\) | \(96\%\) | \(98.35\%\) | \(99.08\%\) | \(99.69\%\) |
SEN | \(96.1\%\) | \(98\%\) | \(98\%\) | \(98.40\%\) | \(99.13\%\) | \(99.74\%\) |
SPE | \(100\%\) | \(99\%\) | \(94.2\%\) | \(93.06\%\) | \(93.88\%\) | \(94.69\%\) |
PRE | \(100\%\) | \(99\%\) | \(94.2\%\) | \(99.93\%\) | \(99.94\%\) | \(99.95\%\) |
F1-score | \(98.01\%\) | \(98.49\%\) | \(96.06\%\) | \(99.16\%\) | \(99.53\%\) | \(99.84\%\) |