Table 3 Comparing average AUPRC for all six data sets over the following methods: miPIE classification tool (restricted to 32 training exemplars), miRDeep2, active learning alone, and dual stage semi-supervised pipeline. Means ± standard deviations are shown, representing 100 repetitions of each experiment (except for miRDeep2)
From: A semi-supervised machine learning framework for microRNA classification
Data set | miPIE | miRDeep2 | Active learning alone | Proposed dual-stage SS pipeline |
|---|---|---|---|---|
hsa | 0.844 (± 0.01) | 0.736 | 0.875 (± 0.01) | 0.903 (± 0.02) |
mmu | 0.966 (± 0.01) | 0.915 | 0.972 (± 0.00) | 0.972 (± 0.00) |
dme | 0.894 (± 0.01) | 0.914 | 0.924 (± 0.01) | 0.941 (± 0.01) |
bta | 0.905 (± 0.02) | 0.869 | 0.935 (± 0.01) | 0.948 (± 0.01) |
gga | 0.919 (± 0.01) | 0.923 | 0.944 (± 0.01) | 0.965 (± 0.00) |
eca | 0.919 (± 0.01) | 0.843 | 0.971 (± 0.00) | 0.971 (± 0.00) |
Average | 0.908 | 0.867 | 0.935 | 0.950 |