mlr3fda (version 0.3.0)

mlr_pipeops_fda.bsignal: B-spline Feature Extraction

Description

This PipeOp extracts features from functional data using B-spline basis functions. The extracted features are B-spline coefficients that represent the functional data in the B-spline basis space. For more details, see FDboost::bsignal(), which is called internally.

Parameters

The parameters are the parameters inherited from PipeOpTaskPreprocSimple, as well as the following parameters:

• inS :: character(1)
  Type of effect in the covariate index: one of "smooth", "linear", "constant". Default "smooth".

• knots :: numeric()
  Either the number of interior knots or a vector of their positions.

• boundary.knots :: numeric(2)
  Boundary points at which to anchor the B-spline basis. Lower and upper boundary points for the spline basis. Defaults to the range of the data.

• degree :: integer(1)
  The degree of the regression spline. Default is 3L.

• differences :: integer(1)
  Order of difference penalty. Default is 1L.

• df :: numeric(1)
  Trace of the hat matrix, controlling smoothness. Default is 4.

• lambda :: any
  Smoothing parameter of the penalty term.

• center :: logical(1)
  Reparameterize the unpenalized part to zero-mean? Default is FALSE.

• cyclic :: logical(1)
  If true the fitted coefficient function coincides at the boundaries.

• Z :: any
  Custom transformation matrix for the spline design.

• penalty :: character(1)
  The penalty type: "ps" (P-spline) or "pss" (shrinkage). DEfault is "ps".

• check.ident :: logical(1)
  Use checks for identifiability of the effect. Default is FALSE.

Super classes

mlr3pipelines::PipeOp -> mlr3pipelines::PipeOpTaskPreproc -> mlr3pipelines::PipeOpTaskPreprocSimple -> PipeOpFDABsignal

Methods

Public methods

• PipeOpFDABsignal$new()

• PipeOpFDABsignal$clone()

Inherited methods

• mlr3pipelines::PipeOp$help()
• mlr3pipelines::PipeOp$predict()
• mlr3pipelines::PipeOp$print()
• mlr3pipelines::PipeOp$train()

---

Method new()

Initializes a new instance of this Class.

Usage

PipeOpFDABsignal$new(id = "fda.bsignal", param_vals = list())

Arguments

id

(character(1))
Identifier of resulting object, default is "fda.bsignal".

param_vals

(named list())
List of hyperparameter settings, overwriting the hyperparameter settings that would otherwise be set during construction. Default list().

---

Method clone()

The objects of this class are cloneable with this method.

Usage

PipeOpFDABsignal$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Examples

task = tsk("fuel")
po_bsignal = po("fda.bsignal")
task_bsignal = po_bsignal$train(list(task))[[1L]]
task_bsignal$data()