Description Usage Arguments Details Value Author(s) Examples

Unscaling the coordinates of points in dimension d

1 2 |

`mat` |
Matrix with values to be unscaled |

`model` |
km object containing knots and eta's. If provided, the arguments allknots and alleta are ignored. |

`allknots` |
List of arrays obtained from the field |

`alleta` |
List of arrays obtained from the field |

`indices` |
Array containing the indices of the dimensions to rescale. |

`standardize` |
If the initial values in the columns of mat are not in [0,S], with S the integral of the piecewise linear function equal to eta at points knots, then there is the possibility to rescale the values y by indicating in which interval they are. |

`lower` |
If |

`upper` |
If |

There are two possible ways to use this function. If the number of columns of mat
is not equal to `d = [email protected]`

(i.e. the number of dimensions), we assume that mat
is allready a submatrix of some bigger matrix with d columns. In that case, all the columns
of mat are rescaled an so `length(indices)`

Must be equal to the number of columns of mat.
The second way to use the function is to have a matrix mat with the same number of columns as d.
In that case only some (or all) columns of mat will be rescaled, depending on the values of indices.
If indices is missing, all the columns are rescaled.

A rescaled matrix with the same size.

Clement Chevalier [email protected]

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 | ```
library(DiceKriging)
myfun <- function(x) return(-1 * branin_robinv(x))
d <- 4
set.seed(8)
n0 <- 40
T <- -10
opt.index <- c(3,4)
inv.index <- c(1,2)
lower <- rep(0,times=d)
upper <- rep(1,times=d)
design <- matrix(runif(d*n0),nrow=n0)
response <- myfun(design)
knots.number <- c(3,0,2,2)
knots <- generate_knots(knots.number = knots.number , d = d)
model <- km(formula = ~1,design = design,response = response,covtype = "matern3_2",scaling = TRUE,knots = knots)
allknots <- model@covariance@knots
alleta <- model@covariance@eta
# generates a sequence of points in dimension 4:
myrands <- matrix( runif(2000),ncol=4 )
result <- unscalingFun(mat = myrands[,1:2], allknots = allknots,alleta = alleta,indices = c(1,2),standardize = TRUE , lower = lower,upper = upper)
result2 <- unscalingFun(mat = myrands, allknots = allknots,alleta = alleta,indices = c(1,2),standardize = TRUE , lower = lower,upper = upper)
# result and result2 are the same
``` |

IRSN/RobustInv documentation built on Dec. 8, 2018, 2:17 a.m.

Embedding an R snippet on your website

Add the following code to your website.

For more information on customizing the embed code, read Embedding Snippets.