Classification handout solutions, problem 2c

[francisco]

I reviewed the solution given, and I think it is wrong, or else I'm
totally lost.

The problem asks to use naive bayes to classify a ghost with a hat, no
moustache, and no bowtie. i.e. (h=1,m=0,b=0)

To calculate, the guess = argmax_y { p(y)*p(m=0|y)*p(b=0|y)*p(h=1|y)}

but, in the calculations they show, they use the probability of p(h=0|y),
which really makes no sense. Is this just an oversight?

Thanks (again),
Francisco

p.s. Daniel, you get the award for most helpful newsgrouper. If I ever see
you at a bar, I'm buying you a beer (which you can totally redeem after
this final).