Table 3 Descriptive statistics for the situations to be considered for assessing the behavior of the best performing machine-learning model
Situation | Classes | Geometric mean | Arithmetic mean | GA-Ratio |
|---|---|---|---|---|
All - Relation Types | 195 | 13.08 | 238.30 | 0.054 |
All - Superclasses | 5 | 6345.04 | 9293.80 | 0.683 |
Scenario 1 - Relation Types | 178 | 10.19 | 141.72 | 0.072 |
Scenario 1 - Superclasses | 5 | 2866.20 | 5045.40 | 0.568 |
Scenario 2 - Relation Types | 95 | 80.09 | 486.00 | 0.165 |
Scenario 2 - Superclasses | 5 | 6308.33 | 9234.00 | 0.683 |
Scenario 3 - Taxonomic | 193 | 12.45 | 240.77 | 0.052 |
Scenario 3 - Biomedical Non-Symmetric | 146 | 10.57 | 318.28 | 0.033 |
Scenario 3 - Biomedical Symmetric | 193 | 12.89 | 240.77 | 0.054 |
Scenario 3 - Non-Biomedical Non-Symmetric | 67 | 47.58 | 693.56 | 0.069 |
Scenario 3 - Non-Biomedical Symmetric | 186 | 12.24 | 249.83 | 0.049 |