Table 8.

Comparison with other systems for both types of mentions combined^a.

Run	Precision	Recall	F1 score
Our best run, mean (SD)	79.60 (2.2)	83.64 (1.2)	81.63 (0.8)
HITb	91.54	83.72	87.45
EZDI	80.90	83.65	82.25
MUSCc	78.90	83.84	81.30
NTTUd	80.43	80.93	80.68
UFe	79.69	79.20	79.44
N2C2f official median	—g	—	76.59
A1h [28]	65.01	88.92	75.10
A2h [28]	85.07	62.11	71.80

^aNational Natural Language Processing Clinical Challenges median is calculated from all valid runs participating in the original evaluation within the shared task.

^bHIT: Harbin Institute of Technology.

^cMUSC: Medical University of South Carolina.

^dNTTU: National Taitung university.

^eUF: University of Florida.

^fN2C2: National Natural Language Processing Clinical Challenges.

^gNot available.

^hThese are variants of the system described in the cited study.