Generates an object of class "RegTypeFamily"
.
L2RegTypeFamily(name, distribution = LMCondDistribution(), distrSymm,
main = 0, nuisance, trafo, param, props = character(0),
L2deriv = EuclRandVarList(EuclRandVariable(
Map = list(function(x) {x[1] * x[2]}),
Domain = EuclideanSpace(dimension = 2),
dimension = 1)),
ErrorDistr = Norm(), ErrorSymm, RegDistr = Norm(), RegSymm,
Regressor = RealRandVariable(Map = list(function(x) {x}), Domain = Reals()),
ErrorL2deriv = EuclRandVarList(RealRandVariable(Map = list(function(x) {x}),
Domain = Reals())),
ErrorL2derivSymm, ErrorL2derivDistr, ErrorL2derivDistrSymm, FisherInfo)
name of the family
conditional distribution (given the regressor)
symmetry of distribution
error distribution
object of class "DistributionSymmetry"
:
symmetry of ErrorDistr
main parameter
optional nuisance parameter
matrix: optional transformation of the parameter
parameter of the family
properties of the family
regressor distribution
object of class "DistributionSymmetry"
:
symmetry of RegDistr
regressor
object of class "EuclRandVariable"
: L2 derivative
object of class "EuclRandVariable"
:
L2 derivative of ErrorDistr
distribution of ErrorL2deriv
object of class "FunSymmList"
:
symmetry of ErrorL2deriv
object of class "DistrSymmList"
:
symmetry of ErrorL2derivDistr
Fisher information matrix
Object of class "L2RegTypeFamily"
If name
is missing, the default
“L2 differentiable regression type family” is used.
If param
is missing, the parameter is created via
main
, nuisance
and trafo
as described
in ParamFamParameter
. In case distrSymm
,
ErrorSymm
, RegSymm
is missing, they are
set to NoSymmetry()
. If FisherInfo
is missing,
it is computed via numerical integration.
Kohl, M. (2005) Numerical Contributions to the Asymptotic Theory of Robustness. Bayreuth: Dissertation.
# NOT RUN {
L2RegTypeFamily()
# }
Run the code above in your browser using DataLab