Table 5 P-values for pairwise comparisons based on the RMSE metric on the phenotypic features over the test set.

From: Powdery mildew resistance prediction in Barley (Hordeum Vulgare L) with emphasis on machine learning approaches

 

RRF

RGP

RNET

RDT

MRF

MGP

MNET

MDT

FRF

FGP

FNET

FDT

RRF

1

0.998

0.994

0.386

1.000

0.103

0.340

0.995

0.879

1.000

0.014

0.459

RGP

0.998

1

1.000

0.035

0.988

0.004

0.931

0.655

1.000

0.963

0.227

0.049

RNET

0.994

1.000

1

0.023

0.973

0.003

0.963

0.563

1.000

0.931

0.296

0.033

RDT

0.386

0.035

0.023

1

0.547

1.000

 < 0.0001

0.969

0.003

0.680

 < 0.0001

1.000

MRF

1.000

0.988

0.973

0.547

1

0.182

0.214

0.999

0.756

1.000

0.006

0.625

MGP

0.103

0.004

0.003

1.000

0.182

1

 < 0.0001

0.714

0.000

0.272

 < 0.0001

1.000

MNET

0.340

0.931

0.963

 < 0.0001

0.214

 < 0.0001

1

0.021

1.000

0.138

0.990

0.000

MDT

0.995

0.655

0.563

0.969

0.999

0.714

0.021

1

0.210

1.000

0.000

0.984

FRF

0.879

1.000

1.000

0.003

0.756

0.000

1.000

0.210

1

0.630

0.680

0.005

FGP

1.000

0.963

0.931

0.680

1.000

0.272

0.138

1.000

0.630

1

0.003

0.752

FNET

0.014

0.227

0.296

 < 0.0001

0.006

 < 0.0001

0.990

0.000

0.680

0.003

1

 < 0.0001

FDT

0.459

0.049

0.033

1.000

0.625

1.000

0.000

0.984

0.005

0.752

 < 0.0001

1

  1. The compared models are RRF (RReliefF-RF), RGP (RReliefF-GP), RNET (RReliefF-NET), RDT (RReliefF-DT), MRF (MRMR-RF), MGP (MRMR-GP), MNET (MRMR-NET), MDT (MRMR-DT), FRF (FTest-RF), FGP (FTest-GP), FNET (FTest-NET), and FDT (FTest-DT).
  2. Significant values are given in bold.
Back to article page