From: Extract antibody and antigen names from biomedical literature
Characteristics | Training | Developing | Testing | Total |
---|---|---|---|---|
No. of PubMed article abstracts | 1930 | 640 | 640 | 3210 |
No. of antibody mentions | 6627 | 1948 | 2260 | 10,835 |
No. of unique antibody mentions | 1963 | 666 | 796 | 3144 |
No. of antigen mentions | 12,198 | 3981 | 4047 | 20,226 |
No. of unique antigen mentions | 3235 | 1335 | 1271 | 4950 |
Avg. sentences per abstract | 7.5 | 7.5 | 7.5 | 7.5 |
Avg. words per abstract | 182.9 | 181.8 | 183.5 | 182.8 |
Avg. mentions per abstract | 9.8 | 9.3 | 9.9 | 9.7 |