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

Table 9. Actual classification performance for TFTF dataset using KNN classifier.

Sensitivity Specificity GMean PPV F1-measure
DWFS 78.23% (±6.39) 57.12% (±6.51) 66.82% (±6.34) 64.79% (±7.13) 70.86% (±6.9)
MRMR+DWFS 80.68% (±5.59) 57.53% (±7.49) 67.97% (±5.07) 65.87% (±5.05) 72.46% (±4.79)
JMI+DWFS 76.18% (±2.41) 60.34% (±9.12) 67.66% (±5.66) 66.39% (±4.12) 70.9% (±2.81)
MRMR 74.96% (±4.83) 57.6% (±9.61) 65.48% (±5.72) 64.35% (±7.18) 69.02% (±4.83)
JMI 76.29% (±5.13) 56.9% (±6.39) 65.83% (±5.2) 64.11% (±6.68) 69.6% (±5.65)
WEKA 77.46% (±6.4) 56.57% (±5.77) 66.1% (±4.91) 64.31% (±5.14) 70.13% (±4.57)
FST3 78.75% (±2.22) 58.38% (±5.02) 67.76% (±3.38) 65.66% (±5.29) 71.53% (±3.65)
ALL Features 75.98% (±5.78) 56.35% (±6.79) 65.36% (±5.53) 63.71% (±7.1) 69.22% (±6.09)
Correlation-baseline 76.23% (±4.36) 55.14% (±7.5) 64.74% (±5.53) 63.25% (±7) 69.05% (±5.7)