Skip to main content

Table 2 Evaluation results on development (D) and test (T) data sets

From: Text mining facilitates database curation - extraction of mutation-disease associations from Bio-medical literature

Sys Set PMD    PM    MD    PD   
   P R F P R F P R F P R F
S1 D 60.0 69.3 64.3 69.2 68.4 68.8 61.7 67.7 64.6 67.2 80.6 73.3
  T 52.6 72.0 60.8 65.5 71.4 68.3 57.0 70.9 63.2 61.1 80.9 69.6
S2 D 76.2 39.0 51.6 82.6 48.1 48.1 74.8 44.3 55.6 84.6 59.8 70.0
  T 77.3 41.3 53.8 76.3 43.0 55.0 67.8 43.6 53.1 74.7 61.4 67.4
S3 D 77.1 36.3 49.7 84.4 45.6 59.2 78.2 42.5 55.0 89.3 57.4 69.9
  T 78.7 36.4 49.7 79.1 38.9 52.2 77.2 41.8 54.3 76.7 59.7 67.2
S4 D 76.4 52.3 60.3 84.4 45.6 59.2 78.2 42.5 55.0 89.3 57.4 69.9
  T 75.8 52.3 61.9 79.1 38.9 52.2 77.2 41.8 54.3 76.7 59.7 67.2
S5 D 75.8 59.6 66.7 81.7 57.0 67.2 75.9 60.0 67.0 88.6 63.8 74.2
  T 71.6 58.3 64.3 76.8 58.0 67.2 74.8 59.3 66.1 76.2 67.7 71.7
  1. Sys – Systems; PMD – Protein-Mutation-Disease relationships; PM – Protein-Mutation relationships; MD – Mutation-Disease relationships; PD – Protein-disease relationships; S1 (System1) – Abstract level co-occurrence; S2 (System2) – Sentence level co-occurrence; S3 (System3) – Sentence level dependency graph based traversal; S4 (System4) – Linking two dependency graphs based on entity identity; S5 (System5) – Linking two or more graphs based on anaphora resolution/trigger words.; P – Precision (in %); R – Recall (in %); F – F-measure (in %)