Table 5 The performance comparison of different ensemble learning methods on better-performing models on the standard images and occluded bird object images test dataset.
Approach | Testing dataset with standard bird objects | Testing dataset with occluded bird objects | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
A (%) | P (%) | R (%) | FS (%) | ACS (%) | A (%) | P (%) | R (%) | FS (%) | ACS (%) | |
Simple Ensemble Learning | 97.56 | 97.56 | 97.6 | 97.58 | 98.51 | 94.77 | 94.16 | 94.59 | 94.37 | 94.68 |
Fuzzy-Based Ensemble Learning | 98.73 | 98.82 | 98.68 | 98.75 | 99.21 | 95.78 | 95.41 | 94.79 | 95.1 | 95.6 |
Random Forest Ensemble Learning | 96.11 | 96.22 | 96.04 | 96.05 | 98.12 | 90.51 | 88.49 | 91.83 | 88.70 | 95.60 |
XG-Boost Ensemble Learning | 89.68 | 91.22 | 89.55 | 89.71 | 98.21 | 75.95 | 73.32 | 79.02 | 71.13 | 95.6 |
