Table 1.
F-measures for each bin and method
| Bin | Exact match | Fingerprint | Bi-gram fingerprint | Combined |
|---|---|---|---|---|
| 1 |
0.80 |
0.98 |
0.99 |
0.99 |
| 2 |
0.83 |
0.92 |
0.94 |
0.96 |
| 3 |
0.86 |
0.97 |
0.98 |
0.99 |
| 4 |
0.80 |
0.96 |
0.95 |
0.99 |
| 5 |
0.85 |
0.96 |
0.98 |
0.98 |
| 6 |
0.88 |
0.94 |
0.97 |
0.97 |
| 7 |
0.88 |
0.97 |
0.99 |
0.99 |
| 8 |
0.92 |
0.96 |
0.97 |
0.97 |
| 9 |
0.86 |
0.97 |
0.97 |
0.98 |
| 10 |
0.93 |
0.98 |
0.98 |
1.00 |
| 11 |
0.90 |
0.97 |
0.98 |
0.99 |
| 12 | 0.91 | 0.98 | 0.98 | 0.99 |
Each F-measure expresses the accuracy of the string match. Note that all the data have been obtained using the UMLS-based term normalisation pre-processes, and these do not indicate how many LFs have links to their corresponding DBpedia titles.