`MLPKriging::copy`

Description

Create a deep copy of an MLPKriging model.

Usage

Python
```
# mk = MLPKriging(...)
mk2 = mk.copy()
```

R

# mk <- MLPKriging(...)
mk2 <- mk$copy()

Matlab/Octave
```
% mk = MLPKriging(...)
mk2 = mk.copy()
```
Julia
```
# mk = MLPKriging(...)
mk2 = copy(mk)
```

Arguments

Argument	Description
None	`copy()` duplicates the current model object without additional arguments.

Details

The returned model is independent from the original one, including the fitted MLP parameters and GP hyper-parameters.

Value

A new independent MLPKriging object with the same fitted parameters.

Examples

f <- function(x) 1 - 1 / 2 * (sin(12 * x) / (1 + x) + 2 * cos(7 * x) * x^5 + 0.7)
X <- as.matrix(seq(0.05, 0.95, length.out = 10))
y <- f(X)

mk <- MLPKriging(
  y, X,
  hidden_dims = c(4L),
  d_out = 1L,
  activation = "tanh",
  kernel = "gauss",
  parameters = list(max_iter_adam = "20", max_iter_bfgs = "10")
)
print(mk)
print(mk$copy())

Results

* MLPKriging
* data: 10x[0.05,0.95] -> 10x[0.163421,0.976851]
* trend constant (est.): -745.4
* variance (est.): 3.13668e+08
* covariance:
  * kernel: gauss
  * range (est.): 2.88992
  * warpings:
      joint: "mlp_joint(4,1,tanh)"  →  MLPJoint(1 -> 4 -> 1, 13 params)
  * total warp params: 13
  * fit:
    * objective: LL
    * optim: BFGS+Adam
* MLPKriging
* data: 10x[0.05,0.95] -> 10x[0.163421,0.976851]
* trend constant (est.): -745.4
* variance (est.): 3.13668e+08
* covariance:
  * kernel: gauss
  * range (est.): 2.88992
  * warpings:
      joint: "mlp_joint(4,1,tanh)"  →  MLPJoint(1 -> 4 -> 1, 13 params)
  * total warp params: 13
  * fit:
    * objective: LL
    * optim: BFGS+Adam