funcc_biclust {FunCC}

Functional Cheng and Church algorithm

Description

The funCC algorithm allows to simultaneously cluster the rows and the columns of a data matrix where each entry of the matrix is a function or a time series

Usage

funcc_biclust(
  fun_mat,
  delta,
  theta = 1,
  template.type = "mean",
  number = 100,
  alpha = 0,
  beta = 0,
  const_alpha = FALSE,
  const_beta = FALSE,
  shift.alignement = FALSE,
  shift.max = 0.1,
  max.iter.align = 100
)

Arguments

fun_mat

The data array (n x m x T) where each entry corresponds to the measure of one observation i, i=1,...,n, for a functional variable m, m=1,...,p, at point t, t=1,...,T

delta

scalar: Maximum of accepted score, should be a real value > 0

theta

scalar: Scaling factor should be a real value > 1

template.type

character: type of template required. If template.type='mean' the template is evaluated as the average function, if template.type='medoid' the template is evaluated as the medoid function.

number

integer: Maximum number of iteration

alpha

binary: if alpha=1 row shift is allowed, if alpha=0 row shift is avoided

beta

binary: if beta=1 row shift is allowed, if beta=0 row shift is avoided

const_alpha

logicol: Indicates if row shift is contrained as constant.

const_beta

logicol: Indicates if col shift is contrained as constant.

shift.alignement

logicol: If shift.alignement=True the shift aligment is performed, if shift.alignement=False no alignment is performed

shift.max

scalar: shift.max controls the maximal allowed shift, at each iteration, in the alignment procedure with respect to the range of curve domains. t.max must be such that 0<shift.max<1

max.iter.align

integer: maximum number of iteration in the alignment procedure

Value

a list of two elements containing respectively the Biclustresults and a dataframe containing the parameters setting of the algorithm @examples data("funCCdata") res <- funcc_biclust(funCCdata,delta=10,theta=1,alpha=1,beta=0,const_alpha=TRUE) res