Table 5 Average values after ten runs on test sets with different fault proportion.
From: A novel hybrid extreme learning machine-based diagnosis model for sensor node faults in aquaculture
algorithm | 5\(\%\) | 10\(\%\) | 15\(\%\) | 20\(\%\) | average value |
|---|---|---|---|---|---|
PSO-KELM | 97.76\(\%\) | 96.25\(\%\) | 95.12\(\%\) | 93.84\(\%\) | 95.74\(\%\) |
PSO-HKELM | 98.10\(\%\) | 98.70\(\%\) | 96.36\(\%\) | 95.90\(\%\) | 97.27\(\%\) |
SSA-HKELM | 99.25\(\%\) | 98.74\(\%\) | 98.83\(\%\) | 98.24\(\%\) | 97.55\(\%\) |
PSO-PNN | 99.30\(\%\) | 99.02\(\%\) | 98.94\(\%\) | 95.46\(\%\) | 98.18\(\%\) |
PSO-CNN | 99.68\(\%\) | 98.90\(\%\) | 98.85\(\%\) | 98.40\(\%\) | 98.88\(\%\) |
UPPSO-HKELM | 99.76\(\%\) | 99.62\(\%\) | 98.75\(\%\) | 99.16\(\%\) | 99.30\(\%\) |
