Table 4 Breakdown of the tokens that were incorrectly tagged as names by our system
From: A system for de-identifying medical message board text
Set | People | Places/ Institutions | Medical | Other | Total N |
|---|---|---|---|---|---|
DEV | 35.3% | 13.5% | 6.3% | 44.9% | 207 |
TEST | 32.9% | 18.8% | 2.9% | 42.0% | 200 |