Skip to main content

Table 9 Comparison of F-score in similar tasks (corpora in brackets); results from [22] referenced from [114]

From: Hybrid natural language processing tool for semantic annotation of medical texts in Spanish

 

UMLS & medical entities

Negation/speculation

[4]

0.715

0.943 (negation), 0.859 (status)

[22]

0.70 [i2b2]

0.63 (negation)

[5]

0.795

0.984 (negation)

[124]

0.855 [i2b2]

0.905 [i2b2] (negation)

 

0.874 [In-house corpus]

0.899 [In-house corpus] (negation)

[61]

 

0.955 (neg. cue), 0.89 (neg. scope)

[NUBEs]

 

0.829 (spec. cue), 0.746 (spec. scope)

[62]

 

0.85 (negation)

[95]

0.798 [Spanish CWLC]

 

[63]

 

0.963 & 0.975 (cue, French)

  

0.765 & 0.880 (scope, French)

[116]

0.658 [Chia] & 0.785 [FRD]

 

[64]

 

0.95 (neg. cue), 0.92 (neg. scope)

[NUBEs]

 

0.84 (spec. cue), 0.80 (spec. scope)

[64]

 

0.90 (neg. cue), 0.84 (neg. scope)

[In-house cancer corpus]

 

0.81 (spec. cue), 0.74 (spec. scope)

[65]

 

0.786 (negation, best model)

[DIANN corpus]

 

0.765 (negation, authors’ model)

[125]

0.660 (disorders)

 

[E3C corpus, Spanish]

  
 

Temporal entities

Medication information

[55]

0.726 [THYME] & 0.762 [i2b2]

 

[120]

0.838 [i2b2]

 

[82]

0.889 [Spanish BARR]

 

[121]

0.824 (scope), 0.783 (class) [HourGlass]

 
 

0.851 (scope), 0.831 (class) [TempEval2]

 

[83]

0.761 (strict), 0.912 (relaxed) [E3C]

 

[126]

0.986 (Age), 0.934 (Frequency), 0.991 (Date)

0.970 (Dose)

[In-house cancer corpus]

0.898 (Duration), 0.933 (Implicit date)

 

[122]

 

0.91–0.95

[66]

 

0.899

[67]

 

0.921 [n2c2 task 2]

 

Miscellaneous entities

Experiencer/event temp.

[60]

 

0.760 (Historical);

[inhouse corpus]

 

1.000 (Experiencer)

[77]

0.800 (Finding) [SpRadIE corpus]

 

[95]

0.622 (Finding)

0.918 (Family_member) [CWLC]

[117]

0.267 (Observation); 0.462 (Qualifier);

 
 

0.735 (Value) [Chia]

 

[123]

0.747 (Observation); 0.618 (Modifier)

0.416 (Family_member) [LEAF]

 

0.967 (Value) [Chia]

Â