For preselected probes, models using different kernel matrices were evaluated as well; again, the bandwidth parameter for the Gaussian kernel was determined based on a grid search via LOO CV.
Figure 17 shows the bandwidth parameter α1 for the displacement and the acceleration as a function of damping ratio.
Low accuracies due to poor selection of the bandwidth parameter, h, for the Gaussian kernel have been documented previously.
Average of diagonal k(x i,x i ) and off-diagonal k(x i,x j ) kernel elements, predictive correlation, and mean-squared error of prediction (MSE) for the diffusion, Gaussian, and two additive genomic relationship kernels at different values of the bandwidth parameter θ for the wheat data.
Averages of diagonal k(x i,x i ) and off-diagonal k(x i,x j ) kernel elements, predictive correlation, and mean-squared error of prediction (MSE) for the diffusion, Gaussian, and two additive genomic relationship kernels (G1∗ and G2∗) with different values of the bandwidth parameter θ for the Holstein data.
Bandwidth parameters and the binaural power k were as used in the cat.
In addition, the bandwidth parameters can be adjusted for the different sizes of kernel functions, which provide a flexible way to define the scale according to research objectives.
The sampling frequency of the actual process will effect an estimated ρ as well as the decision of the bandwidth parameter W. For example, reducing the sample frequency by a factor of 2 will give half the number of data values N, which will increase the in-band bias by a factor of 4, but the reduced number of samples will be fully compensated by the decrease of ρ.
The bandwidth parameter is the same for both SC and OFDM systems.
