R/harmCut.r

In Funclustering: Functional Data Clustering

#
# This function is used in mfpca to cut the harmonic coefficient matrix. In fact
# mfpca call a c++ mfpca which return (among others) the matrix of the coefficients of the harmonics 
# in the basis expantion. But in the multivariate case the coefficient matrix for all dimension
# are store in the same matrix, so we use this function to store each dimension in the specific matrix.     
# 
# @title Separates the matrices of the coefficients of harmonics
# 
# @param harms the coefficients matrix, see the description for more details
# @param nbasis vector containing the number of basis for each dimension
# 
# @return a list of length dimension of the data which contain the harmonic coefficients matrix
# for each dimension
# 
harmsCut <- function(harms, nbasis){
	dimFd=length(nbasis)
	i=0
	iter=1
	harmsCut=list()
	while (iter <= dimFd){
		start=i+1;
		end=i+nbasis[[iter]]
	 	harmsCut[[iter]]=harms[start:end,]
	 	i=end
	 	iter=iter+1
	}
	return(harmsCut)
}