How can a unified confidence score be achieved using multiple reviews?

24 Views Asked by At

For a product P1, there are multiple reviewers: R1, ..., Rn (with corresponding reviewer Quality of Q1, ..., Qn). How should I mathematically combine the Scores (S1, ..., Sn) provided by all the Reviewers to come up with a final quality score?

Note: Reviewer ratings and Quality What I have so far: Final Quality Score = Sum(Qi * Si)/n

Is this correct? Are there other approaches to doing this?

It appears to me that Ensemble learning is the way to go for such a problem. I don't know how it works exactly, so, if anyone has better alternatives, please recommend.