Skip to main content

Table 7 Comparison of experimental results of five kinds of entities recognition of our method and the segment-level neural network

From: Named entity recognition for Chinese judgment documents based on BiLSTM and CRF

Methods Entity type Precision Recall F1
Our method Person name 77.08 73.69 75.35
  Organization name 66.69 61.48 64.10
  Laws and regulation 92.59 94.34 93.46
  Accusation 76.67 71.43 73.95
  Penalty 77.12 72.80 74.90
  Overall 77.08 73.69 75.35
Segment-level neural network Person name 80.54 84.08 82.27
  Organization name 62.55 69.68 65.92
  Laws and regulation 15.54 26.14 19.49
  Accusation 82.12 75.12 78.47
  Penalty 35.96 47.29 40.85
  Overall 71.60 69.40 70.48