Table 2 The table shows the performance comparison (excluding RF) of classifier ensembles derived from different pruning methods and the single best classifier. See Table 1 for more details. Significance levels are as follows: ** p ≤ 0.001, * p ≤ 0.01, and p ≤ 0.05
From: Unsupervised encoding selection through ensemble pruning for biomedical classification
| Â | best | chull | mvo | pfront | rand | single |
|---|---|---|---|---|---|---|
acp_mlacp | 0.69 (±0.1) | 0.68 (±0.1) | 0.82ns (±0.03) | 0.7 (±0.09) | 0.7 (±0.09) | 0.68 (±0.11) |
aip_antiinflam | 0.47 (±0.06) | 0.47 (±0.07) | 0.44 (±0.05) | 0.48** (±0.06) | 0.44 (±0.07) | 0.46 (±0.06) |
amp_antibp2 | 0.89 (±0.04) | 0.88 (±0.03) | 0.9ns (±0.01) | 0.89 (±0.03) | 0.89 (±0.03) | 0.87 (±0.04) |
atb_antitbp | 0.69 (±0.11) | 0.68 (±0.12) | 0.69 (±0.12) | 0.7ns (±0.11) | 0.68 (±0.1) | 0.7 (±0.11) |
avp_amppred | 0.75 (±0.05) | 0.75 (±0.05) | 0.75 (±0.03) | 0.78** (±0.05) | 0.77 (±0.05) | 0.74 (±0.05) |
cpp_mlcpp-complete | 0.74 (±0.05) | 0.74 (±0.05) | 0.75. (±0.06) | 0.75 (±0.05) | 0.75 (±0.05) | 0.72 (±0.05) |
hem_hemopi | 0.87 (±0.05) | 0.87 (±0.05) | 0.84 (±0.04) | 0.88* (±0.05) | 0.88 (±0.05) | 0.87 (±0.05) |
isp_il10pred | 0.56 (±0.08) | 0.52 (±0.09) | 0.56 (±0.06) | 0.56 (±0.08) | 0.52 (±0.07) | 0.58ns (±0.08) |
nep_neuropipred | 0.79 (±0.05) | 0.78 (±0.05) | 0.8. (±0.05) | 0.79 (±0.04) | 0.77 (±0.04) | 0.79 (±0.04) |
pip_pipel | 0.46 (±0.05) | 0.45 (±0.06) | 0.42 (±0.02) | 0.47ns (±0.06) | 0.45 (±0.05) | 0.46 (±0.04) |