Skip to main content

Advertisement

Table 6 Comparison of annotators trained an tested on the filtered corpus described in "Dataset" section

From: CRFVoter: gene and protein related object recognition using a conglomerate of CRF-based tools

System P R F
Stanford NER 0,77 0,69 0,73
MarMoT 0,76 0,69 0,72
CRF++ 0,75 0,64 0,69
MITIE 0,74 0,58 0,65
Glample 0,78 0,72 0,74
Majority vote 0,64 0,72 0,68
CRFVoter 0,75 0,77 0,76