
"funcyCtrl"
Hyperparameters for functional cluster algorithms.
Objects can be created by calls of the form new("funcyCtrl")
. In addition, named lists can be
coerced to funcyCtrl
objects, names are completed if unique (see examples).
Objects of class funcyCtrl
have the following slots:
baseType
:Type of basis functions, one of "eigenbasis"
, "splines"
,
"exponential"
,
"fourier"
,
"power"
,
"polynomial"
.
dimBase
:Dimension of the basis functions.
flexDim
:If TRUE
, dimension can vary between
clusters (if supported by the algorithm). dimBase
is therefore the
maximum dimension.
init
:Algorithm for initial clustering, one of
"kmeans"
, "random"
or "hclust"
.
nrep
:Number of replications for initial clustering.
seed
:Seed number.
thd
:Threshold if fpca was integrated into method.
redDim
:Reduced dimension if coefficients are additionally projected onto lower subspace.
Objects of class funcyCtrlMbc
inherit from
funcyCtrl
and have the following additional slots:
eps
:Convergence threshold for EM-algorithm.
maxit
:Maximum number of iterations.
hard
:Hard classification?
Christina Yassouridis and Dominik Ernst and Friedrich Leisch. Generalization, Combination and Extension of Functional Clustering Algorithms: The R Package funcy. Journal of Statistical Software. 85 (9). 1--25. 2018
# NOT RUN {
##Show slots
showClass("funcyCtrl")
##Define new parameters
mycont = new("funcyCtrl", baseType="fourier", dimBase=4, flexDim=TRUE,
init="hclust")
mycont
# }
Run the code above in your browser using DataLab