Evaluate the log-likelihood of the estimated intrinsic volumes

loglike_ivols evaluates the (normalized) log-likelihood of a vector with respect to given data, the output of prepare_em.

loglike_ivols(v, data, mode = 0)

Arguments

v	vector of mixing weights (conic intrinsic volumes).
data	output of prepare_em(d, m_samp).
mode	specifies whether the first and last values should be taken into account: mode==0: take all into account mode==1: leave out the estimate of the dth intrinsic volume mode==2: leave out the estimate of the 0th intrinsic volume mode==3: leave out both estimates of the 0th and dth intrinsic volume

Value

The output of loglike_ivols is the value of the normalized log-likelihood of the mixing weights v with respect to the sample data given in data

See also

prepare_em, estim_statdim_var

Package: conivol

Examples

D <- c(5,5)
alpha <- c(pi/3,pi/4)
d <- sum(D)
N <- 10^5
v_exact <- circ_ivols( D, alpha, product=TRUE )

# collect sample data
m_samp <- rbichibarsq(N,v_exact)
data <- prepare_em(d, m_samp)
est <- estim_statdim_var(d, m_samp)
v_estim <- list(
    init0 = init_ivols( d ) ,
    init1 = init_ivols( d, 1, delta=est$delta, var=est$var ) ,
    init2 = init_ivols( d, 2, delta=est$delta ) ,
    init3 = init_ivols( d, 3, var=est$var ) ,
    init4 = init_ivols( d, 4, delta=est$delta, var=est$var ) )

# evaluate log-likelihood function
loglike_ivols(v_exact, data)
#> [1] -4.992283
loglike_ivols(v_estim$init0, data)
#> [1] -5.260337
loglike_ivols(v_estim$init1, data)
#> [1] -4.993701
loglike_ivols(v_estim$init2, data)
#> [1] -5.000752
loglike_ivols(v_estim$init3, data)
#> [1] -5.535673
loglike_ivols(v_estim$init4, data)
#> [1] -5.019852