A function to make prediction on robust GaSP models after the robust GaSP model has been constructed.
pred_rgasp(beta, nu, input, X, zero_mean,output, testing_input,
X_testing, L, LX, theta_hat, sigma2_hat,
q_025, q_975, r0, kernel_type, alpha,method,interval_data)inverse-range parameters.
noise-variance ratio parameter.
input matrix.
the mean basis function i.e. the trend function.
The mean basis function is zero or not.
output matrix.
testing input matrix.
mean/trend matrix of testing inputs.
a lower triangular matrix for the cholesky decomposition of R, the correlation matrix.
a lower triangular matrix for the cholesky decomposition of $X^tR^-1X$.
estimated mean/trend parameters.
estimated variance parameter.
0.025 quantile of t distribution.
0.975 quantile of t distribution.
a matrix of absolute difference between inputs and testing inputs.
type of kernel. matern_3_2 and matern_5_2 are Matern kernel with roughness parameter 3/2 and 5/2 respectively. pow_exp is power exponential kernel with roughness parameter alpha. If pow_exp is to be used, one needs to specify its roughness parameter alpha.
Roughness parameters in the kernel functions.
method of parameter estimation. post_mode means the marginal posterior mode is used for estimation. mle means the maximum likelihood estimation is used. mmle means the maximum marginal likelihood estimation is used. The post_mode is the default method.
a boolean value. If T, the interval of the data will be calculated. If F, the interval of the mean of the data will be calculated.
A list of 4 elements. The first is a vector for predictive mean for testing inputs. The second is a vector for lower quantile for 95% posterior credible interval and the third is the upper quantile for 95% posterior credible interval for these testing inputs. The last is a vector of standard deviation of each testing inputs.
Mengyang Gu. (2016). Robust Uncertainty Quantification and Scalable Computation for Computer Models with Massive Output. Ph.D. thesis. Duke University.