estimate Dirichlet parameters.R

library(MCMCpack)

p = rdirichlet(100,rep(1,4))# create some artificial data


###### Shouldn't have to change anything from this point on

# N is the number of stomachs / scat samples you have
N = dim(p)[1]
# prey is the number of items 
prey = dim(p)[2]

findAlpha = function(a) {
	NLL = -sum(log(ddirichlet(p,a)))
	return(NLL)	
}

o = optim(runif(prey),findAlpha)

# these are the estimates of alpha for the prior
o$par