Skip to main content
. 2015 Feb 26;10(2):e0117988. doi: 10.1371/journal.pone.0117988

Table 6. Actual classification performance for Prostate cancer dataset using KNN classifier.

Sensitivity Specificity GMean PPV F1-measure
DWFS 87.56% (±4.9) 85.33% (±13.44) 86.17% (±7.41) 84.57% (±13.5) 85.59% (±8.06)
MRMR+DWFS 94.42% (±5.24) 92.05% (±7.95) 93.17% (±5.68) 92.64% (±8.23) 93.4% (±5.83)
JMI+DWFS 98.18% (±4.07) 86.47% (±7.9) 92.03% (±4.47) 87.44% (±7.21) 92.3% (±4.21)
MRMR 88.93% (±10.55) 88.41% (±14.91) 87.99% (±5.99) 90.31% (±12.61) 88.54% (±5.78)
JMI 85.8% (±5.23) 84.65% (±11.89) 85.14% (±8.44) 84.39% (±11.83) 84.92% (±8.58)
WEKA 75.83% (±19.09) 83.51% (±17.2) 79.29% (±16.66) 79.85% (±20.81) 77.48% (±19.07)
FST3 84.48% (±9.33) 84.12% (±10.11) 84% (±6.19) 83.63% (±9.01) 83.88% (±8.16)
ALL Features 79.58% (±10.6) 83.11% (±11.12) 81.2% (±9.68) 81.02% (±13.52) 79.97% (±10.86)
Correlation-baseline 92.25% (±7.47) 87.15% (±9.92) 89.44% (±5.41) 86.52% (±11.1) 88.84% (±6.53)