Table 3 |
||||
|
Precision, recall, and f-measure of CTD manual curation |
||||
|
Dataa |
Biocurator 1 |
Biocurator 2 |
Biocurator 3 |
Averagee |
|
|
||||
|
No. articles examined |
53 |
68 |
65 |
62 |
|
Precisionb |
0.90 |
0.97 |
0.86 |
0.91 |
|
Recallc |
0.62 |
0.79 |
0.71 |
0.71 |
|
F1-measured |
0.70 |
0.85 |
0.75 |
0.77 |
|
|
||||
|
aOnly curated chemical-gene interactions were analyzed (disease interactions were not considered); consequently, these numbers differ slightly from those reported in Table 2. bPrecision = Correct interactions for individual biocurator/Total interactions for individual biocurator. cRecall = Correct interactions for individual biocurator/Total correct interactions from all biocurators. dF1-measure is the harmonic mean of precision and recall = (2 × Precision × Recall)/(Precision + Recall). Precision, Recall, and F1-measure were macro-averaged over the set of articles curated by that curator. eCalculated using data from three biocurators. |
||||
|
Wiegers et al. BMC Bioinformatics 2009 10:326 doi:10.1186/1471-2105-10-326 |
||||
