Skip to main content

Table 2 Effect of system components on development data

From: Identifying gene and protein mentions in text using conditional random fields

System

Precision

Recall

F-Measure

A. No Lex, No Feat. Ind.

0.793

0.731

0.761

B. No Lexicons

0.807

0.744

0.774

C. Trigrams

0.811

0.759

0.784

D. Non-gene Lexicons

0.818

0.743

0.778

E. Gene Lexicons

0.812

0.775

0.793

F. All Lexicons

0.817

0.782

0.799

  1. A) System containing no lexicon features and does not use feature induction. B) Same as A, except feature induction is used. C) Same as B, except features using the infrequent trigram lexicon are used. D) Same as B, except features using the non-gene lexicons are used. E) Same as B, except features using the gene lexicon are used. F) Same as B, except features using all lexicons are used.