Description Usage Arguments Details Value Author(s) References See Also Examples
Scales a set of proposed inducing point locations in [0,1]^d to center around a reference location, returning the scaled design
1 | qnormscale(X, mean, sd)
|
X |
a |
mean |
a vector representing the reference location to act as the center of the scaling;
|
sd |
a scalar or vector determining the standard deviation for each dimension of the Gaussian CDF |
This function scales a set of proposed inducing points in [0,1]^d to be
centered and concentrated around a reference location. The proposed inducing
points are interpreted as quantiles of one-dimensional
Gaussian distributions centered at the reference location with the standard deviation
provided by the user. For each dimension qnorm
is invoked to rescale the
inducing points.
a matrix
of the scaled set of inducing points
D. Austin Cole austin.cole8@vt.edu
D.A. Cole, R.B. Christianson, and R.B. Gramacy (2021). Locally Induced Gaussian Processes for Large-Scale Simulation Experiments Statistics and Computing, 31(3), 1-21; preprint on arXiv:2008.12857; https://arxiv.org/abs/2008.12857
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 | ## Generate data and define xmean
X <- matrix(runif(30), ncol=2)
xmean <- c(0.3, 0.4) # doesn't need to be in [0,1]^2
## Scale centered at xmean with different standard deviations
X_scaled1 <- qnormscale(X, mean=xmean, sd=.1)
X_scaled2 <- qnormscale(X, mean=xmean, sd=c(.05,.15))
## View scaled X
plot(X, xlab='X1', ylab='X2')
points(xmean[1], xmean[2], pch=16)
points(X_scaled1, pch=2, col=3, lwd=2)
points(X_scaled2, pch=3, col=4, lwd=2)
legend('topright',legend = c('Original X','xmean', 'Xscaled1', 'Xscaled2'),
pch = c(1,16,2,3), col= c(1,1,3,4), lwd=2, lty=NA)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.