Skip to main content

Table 2 Summary of the results from applying the proposed KIGP to the leukemia dataset.

From: Kernel-imbedded Gaussian processes for disease classification using microarray gene expression data

Performance Measure

Test

CV (3-fold)

LOOCV (fixed genes)

 

PK

GK

LK

PK

GK

LK

PK

GK

LK

ERR #

2/34

1/34

1/34

2/38

1/38

1/38

0/38

0/38

0/38

APP

0.858

0.835

0.923

0.844

0.819

0.875

0.995

0.928

0.993

  1. The columns labeled by "Test" are for the independent tests.
  2. The columns labeled by "CV (3-fold)" are for the rigorous 3-fold CVs (each CV involves all three phases of an KIGP).
  3. The columns labeled by "LOOCV (fixed genes)" are for the loose LOOCVs (each CV only involves the "prediction phase" of an KIGP).