Skip to main content

Table 3 Precision of extracted causalities when comparing to human annotators

From: Extracting health-related causality from twitter messages using natural language processing

 

Strict evaluation

Relax evaluation (exclude hypothetical assertions and negation)

Insomnia

73.81%

88.10%

Stress

82.65%

96.94%

Headache

56.10%

85.37%

Micro-average

74.59%

92.27%