Skip to main content

Advertisement

Table 3 Impact of size of the training set on relation extraction

From: Extraction of potential adverse drug events from medical case reports

  Precision Recall F‐score
N Mean SD Mean SD Mean SD
10 0.58 0.41 0.60 0.44 0.55 0.38
20 0.62 0.36 0.69 0.38 0.64 0.37
50 0.79 0.13 0.87 0.06 0.82 0.09
100 0.81 0.05 0.75 0.08 0.78 0.04
200 0.85 0.07 0.84 0.05 0.84 0.04
500 0.82 0.04 0.85 0.01 0.84 0.02
1000 0.83 0.02 0.87 0.02 0.86 0.01
2000 0.87 0.01 0.86 0.01 0.87 0.01
  1. Impact of size of the training set on relation extraction wasmeasured by independent cross‐validations over subsets of theADE‐EXT‐TRAIN corpus. N indicates number of documents inthe training set and SD indicates the standard deviation measuredduring the 10‐fold cross validation.