Skip to main content

Table 3 Performance of the KLOSURE system

From: KLOSURE: Closing in on open–ended patient questionnaires with text mining

QuestionTopicClassesMethodFeaturesP (%)R (%)F (%)
Q1condition3MetaMapN/A95.387.691.3
Q2treatment4MetaMapN/A84.961.671.4
Q3changes3naive Bayes881.380.881.0
Q4confidence3best-first decision tree870.167.366.9
Q5stiffness3reduced error pruning tree885.379.675.6
Q6pain3complement naive Bayes1062.858.359.0
Q7other symptoms3naive Bayes583.283.081.0
Q8daily activities3J48 pruned tree1477.372.373.2
Q9other activities3random forest1471.070.270.6
Q10other comments3Stanford Core NLPN/A75.372.373.8