Table 2.
Results for top 10 Risk Factor Track submissions, ranked by micro-averaged F1
| Rank | Team name | Micro Precision |
Micro Recall |
Micro F1 | System |
|---|---|---|---|---|---|
| 1 | NLM | 0.8951 | 0.9625 | 0.9276 | Additional annotation + lexicon and rules + SVMs |
| 2 | Harbin Institute of Technology Shenzhen Graduate School |
0.9106 | 0.9436 | 0.9268 | Ensemble systems + rules |
| 3 | Kaiser Permanente | 0.8972 | 0.9409 | 0.9185 | SVM+ensemble classifier+rules |
| 4 | Linguamatics and Northwestern University |
0.8975 | 0.9375 | 0.9171 | I2E interface + lexicon + existing tools + rules |
| 5 | University of Nottingham |
0.8847 | 0.9488 | 0.9156 | Lexicon + rules + ML (CRF, NB, ME) |
| 6 | The Ohio State University |
0.8907 | 0.9261 | 0.9081 | Lexicon + rules |
| 7 | TMUNSW | 0.8594 | 0.9387 | 0.8973 | Lexicon + rules + ML (CRF and NB) |
| 8 | National Central University |
0.8586 | 0.9256 | 0.8909 | unknown |
| 9 | UNIMAN | 0.8557 | 0.9007 | 0.8776 | Lexicon + rules |
| 10 | University of Utah | 0.8552 | 0.8951 | 0.8747 | Existing tools + regular expressions |