prepare_em_cm
takes a two-column matrix whose rows form
iid samples from a bivariate chi-bar-squared distribution and
prepares the data used in maximum likelihood estimation.
prepare_em_cm(d, low, upp, m_samp)
d | the dimension of the bivariate chi-bar squared distribution. |
---|---|
low | lower bound for |
upp | upper bound for |
m_samp | two-column matrix whose rows from iid samples from a bivariate chi-bar-squared distribution. |
The output of prepare_em_cm
is (low-upp+1)
row matrix whose
k
th row contains the products of the density values of the chi_k^2
and chi_(d-k)^2 distributions evaluated in the sample points;
the row-form of the matrix is more convenient for the computations.
This function works pretty much exactly as prepare_em
from the
conivol
package, the only difference being that the "boundary
cases" k==0,n
do not have to be considered/are ignored.
In the general case this is not needed, but for the curvature
measures this is a useful feature.
prepare_em
,
constr_eigval
,
constr_eigval_to_bcbsq
,
estim_em_cm
Package: symconivol
CM <- curv_meas_exact(4,3)$A[,2] CM <- CM/sum(CM) m_samp <- conivol::rbichibarsq(1e5,CM) str( prepare_em_cm( 15, 1, 9, m_samp ))#> num [1:9, 1:100000] 0.00178 0.00532 0.00972 0.01332 0.01474 ...