Table 1

Results from two-level nested cross-validation

No. of variables in KNN model

Expected test characteristics in another set of data


Accuracy

SEN (%)

SPEC (%)

PPV (%)

NPV (%)


4 variables

94.44

91.67

95.83

95.65

92.00

6 variables

92.36

87.50

95.83

95.45

88.46

8 variables

92.36

87.50

95.83

95.45

88.46

10 variables

92.36

87.50

95.83

95.45

88.46

21 variables

93.05

89.58

95.83

95.56

90.20

31 variables

92.36

87.50

95.83

95.45

88.46

41 variables

92.36

89.58

95.83

95.56

90.20

51 variables

93.06

89.58

95.83

95.56

90.20


K-Nearest Neighborhood model with Euclidean distance measure & 3 neighbor was used; SEN: sensitivity; SPEC: specificity; PPV: Positive Predictive Value; NPV: Negative Predictive Value.

Kibriya et al. BMC Medical Genomics 2011 4:50   doi:10.1186/1755-8794-4-50

Open Data