Tell me more ×
Cross Validated is a question and answer site for statisticians, data analysts, data miners and data visualization experts. It's 100% free, no registration required.
up vote 4 down vote favorite
2

What are some of the best ranking algorithms with inputs as up and down votes?

share|improve this question
2  
I am a bit unclear as to what exactly you are asking and how data analysis enters the picture? Please elaborate. – user28 Sep 11 '10 at 16:37
How is ranking not related to data analysis? You're trying to combine user data (up/down votes) to generate a rank statistic that conveys information. Admittedly, more work on this has been done in machine learning than statistics, but that's on topic, right? – ars Sep 11 '10 at 18:56
@ars It seems on topic but the question was/is a bit unclear to me. An algorithm can mean many different things depending on context. – user28 Sep 11 '10 at 19:47
@Srikant: there was a comment earlier about closing the question which I was responding to. – ars Sep 11 '10 at 22:56
This isn't just a data analysis question, because many users become aware of the voting system and vote strategically. In other words, the kind of measurement is changing the behavior. Economics (political economy and game theory in particular) has a lot of interesting things to say about choosing ranking algorithms. – David James Jan 3 '11 at 2:16

2 Answers

up vote 6 down vote accepted

David MacIver had a nice post on the algorithm he used for ranking programming languages based on relative rankings from different users for various factors:

It's based the Rank Aggregation Methods for the Web paper by Dwork et al.

If you want something similar to this site, then Reddit's code is open source. There was some discussion of their ranking algorithm on Hacker News a while back:

For a quick and simple take on "hotness", consider Evan Miller's algorithm with time decay:

share|improve this answer
up vote 0 down vote

I have listed a number of schemes for solving this problem at Sorting answers, given overvotes and undervotes. I won't duplicate them here; just take a look over there. The most promising ones look like either Bayesian rating or Laplace smoothing.

share|improve this answer

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.