Research Article
A New Feature Extraction Algorithm Based on Entropy Cloud Characteristics of Communication Signals
Table 1
Recognition rate of different algorithms under different SNR.
| | SNR (dB) | 4 | 0 | ā4 | ā8 |
| | Recognition rate of wavelet transform based features (%) | 98 | 84.6 | 74.5 | 62.5 |
| | Recognition rate of cyclic spectrum entropy based features (%) | 87.5 | 78.5 | 67.0 | 59.5 |
| | Recognition rate of entropy based features (%) | 100 | 97.5 | 96.0 | 78.5 |
| | Recognition rate of entropy cloud model based features (%) | 100 | 100 | 100 | 100 |
|
|
