Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

I'm not even sure if what I want to do is a good idea but I figure I'll experiment and see.

I have two predicted ratings in the range of 1-5 based on two different algorithms for predicting movie ratings. I also have two global averages for the given movie and given user.

e.g.

predicted rating 1 = 4.5 predicted rating 2 = 2.5

global mean movie = 3.8 global mean user = 3.6

The two most important values here that should for now be equally weighed is the predicted ratings but I want to have to tend towards the global means a bit as well if that makes any sense.

What mathematical weight average function can I use and how?

This is all experimental so please tell me if you think i'm way off and you have any other ideas. :) My maths is quite rusty.

share|improve this question
add comment

1 Answer 1

up vote 2 down vote accepted

Possibly what you're looking for is an empirical Bayes method.

Here's a bit more on empirical Bayes methods: http://dspace.mit.edu/bitstream/handle/1721.1/45587/18-441Spring-2002/NR/rdonlyres/Mathematics/18-441Statistical-InferenceSpring2002/0636681D-CFBE-4A76-BC80-D02BB2C2BFB2/0/attachment3.pdf

I suspect you'd have to say more than you wrote above before much more could be said here.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.