Table 2 |
|
|
Features for combined performance |
|
|
not(T) |
Team T did not nominate any gene mention that overlaps with this candidate. |
|
nom(T) |
Team T nominated a gene mention that overlaps with this candidate. |
|
Noms(T, S) |
Team T nominated a gene mention that overlaps with this candidate, and that starts before (S = -1), starts after (S = 1), or coincides with the start of this candidate (S = 0). |
|
Nome(T, E) |
Team T nominated a gene mention that overlaps with this candidate, and that ends before (E = -1), ends after (E = 1), or coincides with the end of this candidate (E = 0). |
|
nom(T, S, E) |
Team T nominated a gene mention with S and E as above. |
|
Noms(S) |
Some team nominated a gene mention with S as above. |
|
Nome(E) |
Some team nominated a gene mention with E as above. |
|
word(W) |
Word W occurs in the candidate. |
|
firstword(W) |
Word W is the first word of this candidate. |
|
lastword(W) |
Word W is the last word of this candidate. |
|
context(P, W) |
Word W at position P relative to this candidate. The possible values for P are -2,-1,1,2. |
|
|
|
|
The features generated for each candidate gene mention, based on the submitted runs. |
|
|
Smith et al. Genome Biology 2008 9(Suppl 2):S2 doi:10.1186/gb-2008-9-s2-s2 |
|