Package 'ddsPLS'

Title: Data-Driven Sparse Partial Least Squares
Description: A sparse Partial Least Squares implementation which uses soft-threshold estimation of the covariance matrices and therein introduces sparsity. Number of components and regularization coefficients are automatically set.
Authors: Hadrien Lorenzo
Maintainer: Hadrien Lorenzo <[email protected]>
License: MIT + file LICENSE
Version: 1.2.1
Built: 2025-02-24 05:34:17 UTC
Source: https://github.com/hlorenzo/ddspls

Help Index

C++ implementation of the bootstrap operations

Description

Start the bootstrap operations. Should not be used by user.

Usage

bootstrap_Rcpp(
  U,
  V,
  X,
  Y,
  lambdas,
  lambda_prev,
  R,
  n_B,
  doBoot,
  n,
  p,
  q,
  N_lambdas,
  lambda0
)

Arguments

U

The weights for X part.
V

The weights for Y part.
X

The matrix of X part.
Y

The matrix of X part.
lambdas

The to be tested values for lambda.
lambda_prev

The previously selected values for lambda.
R

The number of components to build.
n_B

The number of bootstrap samples to generate and analyse.
doBoot

Wheteher do bootstrap operations.
n

The number of observations.
p

The number of variables of X part.
q

The number of variables of Y part.
N_lambdas

The number of to be tested values for lambda.
lambda0

The vector of lambda0

C++ wrapper for bootstrap function

Description

The wrapper used to start the bootstrap commands. Not to be used by the user.

Usage

bootstrapWrap(
  U,
  V,
  X,
  Y,
  lambdas,
  lambda_prev,
  R,
  n_B,
  doBoot = TRUE,
  n,
  p,
  q,
  n_lambdas,
  lambda0.
)

Arguments

U

matrix, weights X
V

matrix, weights Y
X

matrix
Y

matrix
lambdas

vector, the to be tested values for lambda
lambda_prev

vector, the previous selected values for lambda
R

integer, the desired number of components
n_B

integer, the number of bootstrap samples required
doBoot

boolean, whether or not perform bootstrap. Used to build the final model (FALSE)
n

integer, the number of observations
p

integer, the number of covariates
q

integer, the number of response variables
n_lambdas

integer, the number of to be tested lambdas
lambda0.

the vector of lambda0

Value

A list

Data-Driven Sparse Partial Least Squares

Description

The main function of the package. It does both start the ddsPLS algorithm, using bootstrap analysis. Also it estimates automatically the number of components and the regularization coefficients. One regularization parameter per component only is needed to select both in x and in y. Build the optimal model, of the class ddsPLS. Among the different parameters, the lambda is the vector of parameters that are tested by the algorithm along each component for each bootstrap sample. The total number of bootstrap samples is fixed by the parameter n_B, for this parameter, the more the merrier, even if costs more in computation time. This gives access to 3 S3 methods (summary.ddsPLS, plot.ddsPLS and predict.ddsPLS).

Usage

ddsPLS(
  X,
  Y,
  criterion = "diffR2Q2",
  doBoot = TRUE,
  LD = FALSE,
  lambdas = NULL,
  n_B = 50,
  n_lambdas = 100,
  lambda_roof = NULL,
  lowQ2 = 0,
  NCORES = 1,
  errorMin = 1e-09,
  verbose = FALSE
)

Arguments

X

matrix, the covariate matrix (n,p).
Y

matrix, the response matrix (n,q).
criterion

character, whether diffR2Q2 to be minimized, default, or Q2 to be maximized.
doBoot

logical, whether performing bootstrap operations, default to TRUE. If equal to FALSE, a model with is built on the parameters lambda and the number of components is the length of this vector. In that context, the parameter n_B is ignored. If equal to TRUE, the ddsPLS algorithm, through bootstrap validation, is started using lambda as a grid and n_B as the total number of bootstrap samples to simulate per component.
LD

Boolean, wether or not consider Low-Dimensional dataset.
lambdas

vector, the to be tested values for lambda. Each value for lambda can be interpreted in terms of correlation allowed in the model. More precisely, a covariate 'x[j]' is not selected if its empirical correlation with all the response variables 'y[1..q]' is below lambda. A response variable 'y[k]' is not selected if its empirical correlation with all the covariates 'x[1..p]' is below lambda. Default to seq(0,1,length.out = 30).
n_B

integer, the number of to be simulated bootstrap samples. Default to 50.
n_lambdas

integer, the number of lambda values. Taken into account only if lambdas is NULL. Default to 100.
lambda_roof

limit value to be considered in the optimization.
lowQ2

real, the minimum value of Q^2_B to accept the current lambda value. Default to 0.0.
NCORES

integer, the number of cores used. Default to 1.
errorMin

real, not to be used.
verbose

boolean, whether to print current results. Defaut to FALSE.

Value

A list with different interesting output describing the built model

See Also

summary.ddsPLS, plot.ddsPLS, predict.ddsPLS

Examples

# n <- 100 ; d <- 2 ; p <- 20 ; q <- 2
# phi <- matrix(rnorm(n*d),n,d)
# a <- rep(1,p/4) ; b <- rep(1,p/2)
# X <- phi%*%matrix(c(1*a,0*a,0*b,
#                     1*a,3*b,0*a),nrow = d,byrow = TRUE) + matrix(rnorm(n*p),n,p)
# Y <- phi%*%matrix(c(1,0,
#                     0,0),nrow = d,byrow = TRUE) + matrix(rnorm(n*q),n,q)
# model_ddsPLS <- ddsPLS(X,Y,verbose=TRUE)

Applet to start ddsPLS

Description

Applet to start ddsPLS

Usage

ddsPLS_App(...)

Arguments

...

Same parameters as ddsPLS

Value

Mainly visual objects, also possible to save plots

C++ code to build models, internal function

Description

Build a ddsPLS model once the bootstrap operations has allowed to find a correct lambda.

Usage

modelddsPLSCpp_Rcpp(U, V, X, Y, lambdas, R, n, p, q, lambda0)

Arguments

U

The weights for X part.
V

The weights for Y part.
X

The matrix of X part.
Y

The matrix of X part.
lambdas

The to be tested values for lambda.
R

The number of components to build.
n

The number of observations.
p

The number of variables of X part.
q

The number of variables of Y part.
lambda0

The vector of regulation parameters.

Function to plot bootstrap performance results of the ddsPLS algorithm

Description

Function to plot bootstrap performance results of the ddsPLS algorithm

Usage

## S3 method for class 'ddsPLS'
plot(
  x,
  type = "criterion",
  digits = 1,
  legend.position = "topright",
  horiz = TRUE,
  biPlot = FALSE,
  las = 0,
  col = NULL,
  cex.names = 1,
  mar = c(5, 4, 4, 2) + 0.1,
  ...
)

Arguments

x

A ddsPLS object
type

The type of graphics. One of "criterion" (default), "total", "prop", "predict", "Q2r", "Q2", "R2r", "R2", "weightX", "weightY", "loadingX" or "loadingY".
digits

double. Rounding of the written explained variance.
legend.position

character. Where to put the legend.
horiz

boolean. Whether to plot horizontally.
biPlot

boolean wether or not to plot one component versus the other.
las

numeric in (0,1,2,3): the style of axis labels.
col

vector. Mainly to modify bars in weight plots.
cex.names

double. Size factor for variable names.
mar

vector. The margins for the plot.
...

Other plotting parameters to affect the plot.

See Also

ddsPLS, predict.ddsPLS, summary.ddsPLS

Function to predict from ddsPLS objects

Description

Function to predict from ddsPLS objects

Usage

## S3 method for class 'ddsPLS'
predict(
  object,
  X_test = NULL,
  toPlot = FALSE,
  doDiagnosis = T,
  legend.position = "topright",
  cex = 1,
  cex.text = 1,
  ...
)

Arguments

object

A ddsPLS object.
X_test

matrix, a test data-set. If is "NULL", the default value, the predicted values for the train test are returned.
toPlot

boolean, wether or not to plot the extreme value test plot. Default to 'TRUE'.
doDiagnosis

Yes or no to perform diagnosis.
legend.position

character. Where to put the legend.
cex

float positive. Number indicating the amount by which plotting symbols should be scaled relative to the default.
cex.text

float positive. Number indicating the amount by which plotting text elements should be scaled relative to the default.
...

Other parameters

See Also

ddsPLS, plot.ddsPLS, summary.ddsPLS

Function to sum up bootstrap performance results of the ddsPLS algorithm

Description

Function to sum up bootstrap performance results of the ddsPLS algorithm

Usage

## S3 method for class 'ddsPLS'
print(x, ...)

Arguments

x

A ddsPLS object.
...

Other parameters to be taken into account.

See Also

ddsPLS, plot.ddsPLS, predict.ddsPLS

Function to sum up bootstrap performance results of the ddsPLS algorithm

Description

Function to sum up bootstrap performance results of the ddsPLS algorithm

Usage

## S3 method for class 'ddsPLS'
summary(
  object,
  return = FALSE,
  plotSelection = FALSE,
  las = 1,
  cex.names = 1,
  digits = 2,
  ...
)

Arguments

object

A ddsPLS object.
return

Wether or not to return the printed values, default to FALSE.
plotSelection

boolean. Whether plot the selection variables.
las

interger. Parameter for angle of variable names.
cex.names

real positive. Which factor zomm the variable names.
digits

integer indicating the number of decimal places (round) to be used.
...

Other parameters to be taken into account.

See Also

ddsPLS, plot.ddsPLS, predict.ddsPLS