Skip to main content

Table 4 Software algorithm measures of performance in test and tenth-round training set

From: Identifying influenza-like illness presentation from unstructured general practice clinical narrative using a text classifier rule-based expert system versus a clinical expert

Measure Training (round 10) Test
Estimated value 95 % Confidence intervals
Lower limit Upper limit
Incidence 0.136 0.124 0.103 0.148
Sensitivity 0.882 0.902 0.831 0.950
Specificity 0.983 0.982 0.970 0.990
Positive Predictive Value 0.893 0.878 0.804 0.930
F-measure 0.888 0.890 0.817 0.940