Skip to main content

Table 3 Mean normalised classification rates over 4 two-phenotype datasets according to performance

From: Pathway activity inference for multiclass disease classification through a mathematical programming optimisation framework

Two-class DIGS SG Mean [ [28]] Median [ [28]] Per_pathway [ [72]] PCA [ [45]] CORGs [ [19]]
5-NN 0.9988 0.9071 0.8737 0.8751 0.8903 0.8389 0.9371
NN 0.9973 0.9584 0.9323 0.9004 0.9041 0.9480 0.9769
SMO 0.9757 0.9474 0.9435 0.9225 0.9325 0.9704 0.9645
HB 0.9835 0.9730 0.8819 0.8707 0.8547 0.8402 0.9595
Logistic 0.9318 0.9816 0.8902 0.8789 0.8632 0.8482 0.9684
Mean 0.9774 0.9535 0.9043 0.8895 0.8890 0.8891 0.9613
  1. The highest classification rate achieved across all competing methods is highlighted as bold for each classifier.