Corr Gauss GP using inherited optim

Corr Gauss GP using inherited optim

Corr Gauss GP using inherited optim

Format

R6Class object.

Value

Object of R6Class with methods for fitting GP model.

Super class

GauPro::GauPro -> GauPro_kernel_model_LOO

Public fields

tmod

A second GP model for the t-values of leave-one-out predictions

use_LOO

Should the leave-one-out error corrections be used?

Methods

Public methods

• GauPro_kernel_model_LOO$new()

• GauPro_kernel_model_LOO$update()

• GauPro_kernel_model_LOO$pred_one_matrix()

• GauPro_kernel_model_LOO$clone()

Inherited methods

• GauPro::GauPro$AIC()
• GauPro::GauPro$AugmentedEI()
• GauPro::GauPro$CorrectedEI()
• GauPro::GauPro$EI()
• GauPro::GauPro$KG()
• GauPro::GauPro$cool1Dplot()
• GauPro::GauPro$deviance()
• GauPro::GauPro$deviance_fngr()
• GauPro::GauPro$deviance_grad()
• GauPro::GauPro$fit()
• GauPro::GauPro$get_optim_functions()
• GauPro::GauPro$grad()
• GauPro::GauPro$grad_dist()
• GauPro::GauPro$grad_norm()
• GauPro::GauPro$grad_norm2_dist()
• GauPro::GauPro$grad_norm2_mean()
• GauPro::GauPro$grad_norm2_sample()
• GauPro::GauPro$grad_sample()
• GauPro::GauPro$gradpredvar()
• GauPro::GauPro$hessian()
• GauPro::GauPro$importance()
• GauPro::GauPro$loglikelihood()
• GauPro::GauPro$maxEI()
• GauPro::GauPro$maxqEI()
• GauPro::GauPro$optim()
• GauPro::GauPro$optimRestart()
• GauPro::GauPro$optimize_fn()
• GauPro::GauPro$param_optim_lower()
• GauPro::GauPro$param_optim_start()
• GauPro::GauPro$param_optim_start0()
• GauPro::GauPro$param_optim_start_mat()
• GauPro::GauPro$param_optim_upper()
• GauPro::GauPro$plot()
• GauPro::GauPro$plot1D()
• GauPro::GauPro$plot2D()
• GauPro::GauPro$plotLOO()
• GauPro::GauPro$plot_track_optim()
• GauPro::GauPro$plotkernel()
• GauPro::GauPro$plotmarginal()
• GauPro::GauPro$plotmarginalrandom()
• GauPro::GauPro$pred()
• GauPro::GauPro$pred_LOO()
• GauPro::GauPro$pred_mean()
• GauPro::GauPro$pred_meanC()
• GauPro::GauPro$pred_var()
• GauPro::GauPro$pred_var_after_adding_points()
• GauPro::GauPro$pred_var_after_adding_points_sep()
• GauPro::GauPro$pred_var_reduction()
• GauPro::GauPro$pred_var_reductions()
• GauPro::GauPro$predict()
• GauPro::GauPro$print()
• GauPro::GauPro$sample()
• GauPro::GauPro$summary()
• GauPro::GauPro$update_K_and_estimates()
• GauPro::GauPro$update_corrparams()
• GauPro::GauPro$update_data()
• GauPro::GauPro$update_fast()
• GauPro::GauPro$update_nugget()
• GauPro::GauPro$update_params()

---

Method new()

Create a kernel model that uses a leave-one-out GP model to fix the standard error predictions.

Usage

GauPro_kernel_model_LOO$new(..., LOO_kernel, LOO_options = list())

Arguments

...

Passed to super$initialize.

LOO_kernel

The kernel that should be used for the leave-one-out model. Shouldn't be too smooth.

LOO_options

Options passed to the leave-one-out model.

---

Method update()

Update the model. Should only give in (Xnew and Znew) or (Xall and Zall).

Usage

GauPro_kernel_model_LOO$update(
  Xnew = NULL,
  Znew = NULL,
  Xall = NULL,
  Zall = NULL,
  restarts = 5,
  param_update = self$param.est,
  nug.update = self$nug.est,
  no_update = FALSE
)

Arguments

Xnew

New X values to add.

Znew

New Z values to add.

Xall

All X values to be used. Will replace existing X.

Zall

All Z values to be used. Will replace existing Z.

restarts

Number of optimization restarts.

param_update

Are the parameters being updated?

nug.update

Is the nugget being updated?

no_update

Are no parameters being updated?

---

Method pred_one_matrix()

Predict for a matrix of points

Usage

GauPro_kernel_model_LOO$pred_one_matrix(
  XX,
  se.fit = F,
  covmat = F,
  return_df = FALSE,
  mean_dist = FALSE
)

Arguments

XX

points to predict at

se.fit

Should standard error be returned?

covmat

Should covariance matrix be returned?

return_df

When returning se.fit, should it be returned in a data frame?

mean_dist

Should mean distribution be returned?

---

Method clone()

The objects of this class are cloneable with this method.

Usage

GauPro_kernel_model_LOO$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Examples

n <- 12
x <- matrix(seq(0,1,length.out = n), ncol=1)
y <- sin(2*pi*x) + rnorm(n,0,1e-1)
gp <- GauPro_kernel_model_LOO$new(X=x, Z=y, kernel=Gaussian)
y <- x^2 * sin(2*pi*x) + rnorm(n,0,1e-3)
gp <- GauPro_kernel_model_LOO$new(X=x, Z=y, kernel=Matern52)
#> * nug is at minimum value after optimizing. Check the fit to see it this caused a bad fit. Consider changing nug.min. This is probably fine for noiseless data.
#> * nug is at minimum value after optimizing. Check the fit to see it this caused a bad fit. Consider changing nug.min. This is probably fine for noiseless data.
y <- exp(-1.4*x)*cos(7*pi*x/2)
gp <- GauPro_kernel_model_LOO$new(X=x, Z=y, kernel=Matern52)
#> * nug is at minimum value after optimizing. Check the fit to see it this caused a bad fit. Consider changing nug.min. This is probably fine for noiseless data.
#> * nug is at minimum value after optimizing. Check the fit to see it this caused a bad fit. Consider changing nug.min. This is probably fine for noiseless data.