SBCK (version 1.0.0)

AR2D2: AR2D2 (Analogues Rank Resampling for Distributions and Dependences) method

Description

Perform a multivariate (non stationary) bias correction.

Public fields

mvq

[MVQuantilesShuffle] Class to transform dependance structure

bc_method

[SBCK::] Bias correction method

bckwargs

[list] List of arguments of bias correction

bcm_

[SBCK::] Instancied bias correction method

reverse

[bool] If we apply bc_method first and then shuffle, or reverse

Methods

Public methods

• AR2D2$new()

• AR2D2$fit()

• AR2D2$predict()

• AR2D2$clone()

---

Method new()

Create a new AR2D2 object.

Usage

AR2D2$new(
  col_cond = base::c(1),
  lag_search = 1,
  lag_keep = 1,
  bc_method = SBCK::CDFt,
  shuffle = "quantile",
  reverse = FALSE,
  ...
)

Arguments

col_cond

Conditionning colum

lag_search

Number of lags to transform the dependence structure

lag_keep

Number of lags to keep

bc_method

Bias correction method

shuffle

Shuffle method used, can be quantile or rank

reverse

If we apply bc_method first and then shuffle, or reverse

...

Others named arguments passed to bc_method$new

Returns

A new `AR2D2` object.

---

Method fit()

Fit the bias correction method. If X1 is NULL, the method is considered as stationary

Usage

AR2D2$fit(Y0, X0, X1 = NULL)

Arguments

Y0

[matrix: n_samples * n_features] Observations in calibration

X0

[matrix: n_samples * n_features] Model in calibration

X1

[matrix: n_samples * n_features] Model in projection

Returns

NULL

---

Method predict()

Predict the correction

Usage

AR2D2$predict(X1 = NULL, X0 = NULL)

Arguments

X1

[matrix: n_samples * n_features or NULL] Model in projection

X0

[matrix: n_samples * n_features or NULL] Model in calibration

Returns

[matrix or list] Return the matrix of correction of X1 if X0 is NULL (and vice-versa), else return a list containing Z1 and Z0, the corrections of X1 and X0

---

Method clone()

The objects of this class are cloneable with this method.

Usage

AR2D2$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Details

Use Quantiles shuffle in calibration and projection period with CDFt

References

Vrac, M. et S. Thao (2020). “R2 D2 v2.0 : accounting for temporal dependences in multivariate bias correction via analogue rank resampling”. In : Geosci. Model Dev. 13.11, p. 5367-5387. doi :10.5194/gmd-13-5367-2020.

Examples

## Three 4-variate random variables
Y0 = matrix( stats::rnorm( n = 1000 ) , ncol = 4 ) ## Biased in calibration period
X0 = matrix( stats::rnorm( n = 1000 ) , ncol = 4 ) / 2 + 3 ## Reference in calibration period
X1 = matrix( stats::rnorm( n = 1000 ) , ncol = 4 ) * 2 + 6 ## Biased in projection period

## Bias correction
cond_col = base::c(2,4)
lag_search = 6
lag_keep = 3
## Step 1 : construction of the class AR2D2 
ar2d2 = SBCK::AR2D2$new( cond_col , lag_search , lag_keep ) 
## Step 2 : Fit the bias correction model
ar2d2$fit( Y0 , X0 , X1 )
## Step 3 : perform the bias correction
Z = ar2d2$predict(X1,X0)