Skip to main content

Table 3 IBT+I vs. task-3 participants (TEES and LIMSI) evaluated on test set by the online system of the BioNLP-ST 2013 task; relations without gold entities; relaxed scores in parenthesis

From: Structured learning for spatial information extraction from biomedical text: bacteria biotopes

System

P

R

F

TEES

0.18 (0.61)

0.12 (0.41)

0.14 (0.49)

LIMSI

0.12 (0.15)

0.04 (0.08)

0.06 (0.09)

IBT+IG1 (p)

0.311

0.171

0.221

IBT+I

0.238 (0.596)

0.279 (0.561)

0.257 (0.578)

IBT+IG1

0.311 (0.594)

0.241 (0.483)

0.272 (0.533)

IBT+IG2

0.331 (0.588)

0.224 (0.431)

0.267 (0.498)

IBT+I (s)

0.241 (0.515)

0.436 (0.624)

0.311 (0.564)

IBT+IG1 (s)

0.305 (0.563)

0.400 (0.640)

0.346 (0.599)

IBT+IG2 (s)

0.327 (0.555)

0.367 (0.560)

0.346 (0.558)

  1. (s) denotes the sentence level evaluation. (p) denotes that the strict evaluation also is punished by missing PartOf relations.