Joint probability mass function maximum likelihood estimator

247 Views Asked by At

I'm stuck on a question finding the maximum likelihood estimators of

$$p(x,y)=\dfrac{n!}{x!\cdot y!\cdot (n-x-y)!\cdot }{p^x\cdot q^y\cdot (1-p-q)^{n-x-y}}$$

I think it is a multinomial distribution but am struggling to get the likelihood function because of the number of variables. How would you get the likelihood function for this?

1

There are 1 best solutions below

0
On

You should change the name of the pmf as it conflicts with the names of the parameters. In this case $$ f(x,y,p,q)=\frac{n!}{x!y!(n-x-y)!}p^yq^y(1-p-q)^{n-x-y}. $$ The Likelihood is the pmf viewed as a function of $p$ and $q$ for fixed $x$ and $y$. In other words $$ L_{x,y}(p,q)=\frac{n!}{x!y!(n-x-y)!}p^yq^y(1-p-q)^{n-x-y}\tag{1}. $$ If you wish to find the log-likelihood $\ell(p,q)$, simply take the logarithm of (1).