Table 1 Single-environment prediction performance of 7 kernels under a ten-fold cross-validation with data Data_Wheat_2019.RData.

From: A guide for kernel generalized regression methods for genomic-enabled prediction

Kernel

MSE

SE_MSE

Cor

SE_Cor

Time

Linear

5.279

0.497

0.642

0.048

10.140

Polynomial

4.974

0.587

0.658

0.054

10.510

Sigmoid

5.594

0.492

0.625

0.049

9.370

Gaussian

4.881

0.540

0.668

0.047

12.440

AK1

5.149

0.522

0.651

0.047

10.600

AKL

4.754

0.515

0.678

0.047

9.970

Exponential

4.980

0.521

0.663

0.048

9.780

  1. MSE denotes the mean square error of prediction (SE_MSE id the standard error of MSE) and Cor is used for the average Pearson’s correlation (SE_Cor is the standard error of Cor). Time indicates the implementation time in seconds.
Back to article page