TABLE 4.
Random Forest predictions on different embeddings.
| With tissue | No tissue | ||||||
|---|---|---|---|---|---|---|---|
| Category | Precision | Recall | F1 | Category | Precision | Recall | F1 |
| DistMult/Cos-sim/@all | 0.5339 | 0.1609 | 0.247278 | DistMult_notissue/Cos-sim/@all | 0.3747 | 0.0326 | 0.059981 |
| DistMult/Euclidean/@all | 0.6758 | 0.2413 | 0.355622 | DistMult_notissue/Euclidean/@all | 0.4152 | 0.0917 | 0.150222 |
| RDF2Vec/Cos-sim/@all | 0.4765 | 0.2057 | 0.287353 | RDF2Vec_notissue/Cos-sim/@all | 0.5711 | 0.1636 | 0.25434 |
| RDF2Vec/Euclidean/@all | 0.412 | 0.242 | 0.304905 | RDF2Vec_notissue/Euclidean/@all | 0.4074 | 0.1246 | 0.190835 |
| TransD/Cos-sim/@all | 0.7356 | 0.3827 | 0.503468 | TransD_notissue/Euclidean/@all | 0.6038 | 0.3066 | 0.40669 |
| TransD/Euclidean/@all | 0.5312 | 0.3462 | 0.419196 | TransD_notissue/Cos-sim/@all | 0.6794 | 0.1027 | 0.178428 |
| TransE/Cos-sim/@all | 0.6988 | 0.6854 | 0.692035 | TransE_notissue/Cos-sim/@all | 0.6894 | 0.6049 | 0.644392 |
| TransE/Euclidean/@all | 0.6604 | 0.5085 | 0.57458 | TransE_notissue/Euclidean/@all | 0.5958 | 0.3098 | 0.407639 |
| TransH/Cos-sim/@all | 0.6922 | 0.5884 | 0.636093 | TransH_notissue/Euclidean/@all | 0.54 | 0.3021 | 0.387446 |
| TransH/Euclidean/@all | 0.6187 | 0.5818 | 0.599683 | TransH_notissue/Cos-sim/@all | 0.6601 | 0.6263 | 0.642756 |
Bold numbers show the highest performance.