Skip to main content

Table 6 Performance by sentences.

From: Mining clinical relationships from patient narratives

   Number of sentence boundaries between arguments
   inter- intra- inter- and intra-sentential
Relation Metric 1n5 n < 1 n1 n2 n3 n4 n5
has_finding P 24 68 65 62 60 61 61
  R 18 89 81 79 78 78 77
  F1 18 76 72 69 67 68 67
has_indication P 18 49 42 42 36 32 30
  R 17 59 47 42 42 39 38
  F1 16 51 42 39 37 34 33
has_location P n/a 74 72 73 72 72 72
  R n/a 83 81 81 81 82 82
  F1 n/a 77 75 76 75 76 76
has_target P 3 64 62 59 60 59 58
  R 1 75 66 64 62 61 61
  F1 2 68 63 61 60 60 59
laterality_modifies P n/a 86 84 86 86 86 87
  R n/a 89 88 88 88 87 88
  F1 n/a 85 84 85 86 85 86
negation_modifies P n/a 80 79 79 80 80 80
  R n/a 94 93 91 93 93 93
  F1 n/a 86 85 84 85 86 85
sub_location_modifies P n/a 89 88 88 89 89 89
  R n/a 95 95 95 95 95 95
  F1 n/a 91 91 91 91 91 91
Overall P 22 69 65 64 62 61 60
  R 17 83 75 73 71 70 70
  F1 19 75 69 68 66 65 65
  1. Variation in performance, by number of sentence boundaries (n) crossed by a relationship. For all cases, the cumulative feature set +event of Table 4 was used. For the inter-sentential-only classifier 1 ≤ n ≤ 5, the score fields for some relations are marked as n/a (not applicable). This is because some relations are either absent from the inter-sentential data (i.e. only ever appear intra-sententially), or are so rare that they do not appear in all training/test folds, and so a macro-average cannot be computed across the folds.
\