Estimate the normalized columns mu of the beta matrix parameter in a mixture of
logistic regressions models, with a spectral method described in the package vignette.
Usage
computeMu(X, Y, optargs = list())
Value
The estimated normalized parameters as columns of a matrix mu of size dxK
Arguments
X
Matrix of input data (size nxd)
Y
Vector of binary outputs (size n)
optargs
List of optional argument:
'jd_method', joint diagonalization method from the package jointDiag:
'uwedge' (default) or 'jedi'.
'jd_nvects', number of random vectors for joint-diagonalization
(or 0 for p=d, canonical basis by default)
'M', moments of order 1,2,3: will be computed if not provided.
'K', number of populations (estimated with rank of M2 if not given)
See Also
multiRun to estimate statistics based on mu,
and generateSampleIO for I/O random generation.
io <- generateSampleIO(10000, 1/2, matrix(c(1,0,0,1),ncol=2), c(0,0), "probit")
mu <- computeMu(io$X, io$Y, list(K=2)) #or just X and Y for estimated K