How to normalize reviews based on evaluation

Question

What is the best way to normalize reviews? I.E. suggests that we have products that users can vote from 1-5 stars.

Just averaging is not a good way, because it does not take into account the number of reviews.

For example, if a product has only one 5 star review, it should not be ahead of a product with 10,000 reviews, simply because a single review gave it 5 stars.

Essentially, how can I normalize an estimate based on the number of reviews?

Answer 1

Sorry if my answer looks crazy. But when I first saw your question, in my opinion the following answer came.

The formula for calculating 250 items gives true Bayesian assessment:

weighted rating (WR) = (v ÷ (v+m)) × R + (m ÷ (v+m)) × C 

Where:

R = average value for the film (average) = (Rating)

v = number of votes per movie = (votes)

m = minimum votes to be listed in the Top 250 (currently 3,000)

C = average vote for the entire report (currently 6.9)

(So ​​IMDB rates its best films according to user reviews and votes. Below is a link to the page where I received the above excerpt: http://www.imdb.com/chart/top .)