set.seed(9)
# We simulate the binomial data with some fixed parameters and
# then try to reach the same estimations.
m <- 10
k <- 100
p <- 0.654
y <- rBI(k,m,p) #Simulations
# without dispersion parameter
BIest(y,m)
# with dispersion parameter
# estimation by method of moments.
BIest(y,m,disp=TRUE,method="MM")
# estimation by maximum quasi-likelihood.
BIest(y,m,disp=TRUE,method="MLE")
Run the code above in your browser using DataLab