5 star ratings. Bayesian or Weighted average?
Similar question, but unlike that question this deals with two factors the stars and reputation.
Lets say Stack Exchange were to have a 5 star rating system instead of up vote and down vote.
- 1 star = 5 reputation
- 2 stars = -2 reputation
- 3 stars = 0 reputation
- 4 stars = +2 reputation
- 5 stars = +5 reputation
So far so good, as it is very easy to calculate the total reputation (just by adding).
But what if each user had a 5 star rate, instead of reputation number? How could we calculate the stars?