# maximum likelihood discriminant rule

I'm trying to use maximum discriminant rule to classify objects in my database. Could, please, somebody clarify the following for me:
For example we have 3 groups and we already know what groups our objects belong to. I want to apply ML discriminant rule to check it's efficiency. (for all these operations I use matlab)

mu = mean(training)
sigma = cov(training)
P = mvnpdf(training, mu, sigma)%multivariant density function for gauss distribution


We doing it for each group, as I understood. What we need to do next to classify objects and what is $\max\limits_{1 \le i \le k}\ L_i (x)$ in discriminant rule which we use to classify objects: $$L_i (x)\ = \max\limits_{1 \le i \le k}\ L_i (x)\\$$ as I understood $L_i (x)$ - value found from the command above P = mvnpdf(training, mu, sigma)

I found this source, explaining the way how rule can be applied to 2 groups, but I do NOT see any connection between theory and example oultined there (totally different things are shown). Moreover I need to apply this rule to classify my objects into 3 groups while in this article they describe 2 groups case.

Any help greatly appreciated!

-