Skip to main content

Advertisement

Table 9 Analysis of LVH versus OVA

From: Improved classification of large imbalanced data sets using rationalized technique: Updated Class Purity Maximization Over_Sampling Technique (UCPMOT)

Data set Over_sampling techniques
B C D E F G H I J K
%RDD %RDA %RDD %RDA %RDD %RDA %RDD %RDA %RDD %RDA %RDD %RDA %RDD %RDA %RDD %RDA %RDD %RDA %RDD %RDA
Yeast 84.4 1.34 87.69 1.3 85.27 1.32 87.51 2.41 88.22 1.34 86.03 2.38 51.47 2.35 23.09 2.32 87.19 1.16 83.44 2.56
Car 11.76 2.12 9.12 1.05 5.91 0 10.37 2.11 5.91 0 9.91 1.04 1.79 1.03 9.59 1.03 11.34 0 6.16 2.12
KEGG-U 4.97 2.19 3.81 2.17 3.28 3.27 4.15 2.15 13.01 3.31 6.31 1.07 3.89 1.06 4.85 1.04 5.38 1.05 6.57 1.08
Average 34.73 1.88 33.54 1.13 31.48 1.53 34.01 2.22 35.71 1.55 34.08 1.49 19.05 1.11 12.51 1.1 34.64 0.55 32.6 1.44
  1. %RDD: % relative difference in data set instances using OVA over LVH method in comparison to initial data set
  2. %RDA: % relative difference in AUC values (Random Forest–classifier) of OVA over LVH method in comparison to base AUC values