R/llik.R
llikNbinomMu.Rd
Calculate the log likelihood of the negative binomial function (and its derivatives)
llikNbinomMu(x, size, mu, full = FALSE)
Number of successes
Size of trial
mu parameter for negative binomial
Add the data frame showing x, mean, sd as well as the fx and derivatives
data frame with fx
for the pdf value of with
dProb
that has the derivatives with respect to the parameters at
the observation time-point
llikNbinomMu(46:54, 100, 40)
#> fx dMu
#> 1 -3.326828 0.1071429
#> 2 -3.446132 0.1250000
#> 3 -3.579663 0.1428571
#> 4 -3.727034 0.1607143
#> 5 -3.887874 0.1785714
#> 6 -4.061827 0.1964286
#> 7 -4.248554 0.2142857
#> 8 -4.447728 0.2321429
#> 9 -4.659037 0.2500000
llikNbinomMu(46:54, 100, 40, TRUE)
#> x size mu fx dMu
#> 1 46 100 40 -3.326828 0.1071429
#> 2 47 100 40 -3.446132 0.1250000
#> 3 48 100 40 -3.579663 0.1428571
#> 4 49 100 40 -3.727034 0.1607143
#> 5 50 100 40 -3.887874 0.1785714
#> 6 51 100 40 -4.061827 0.1964286
#> 7 52 100 40 -4.248554 0.2142857
#> 8 53 100 40 -4.447728 0.2321429
#> 9 54 100 40 -4.659037 0.2500000