Geek Answers Handbook

Best matchmaking algorithm for ranking sources in a crowd?

I would like to create a system in which the crowd will display the top 10 items from the set, which can vary from 20 to 2000 points (ranking among the top ten is not important). There is an excellent stackoverflow article for algorithms to do the actual sorting in How to rank a million images with crowdsourcing type . I am inclined to the requests of users who they like best between the two elements, and then using the TrueSkill algorithm .

My question is asked. I'm using something like TrueSkill, what is the best algorithm for deciding which pairs of elements to show the user? I will have a limited number of opportunities to ask people what subjects they like most, so it is important that the pairs presented provide the system with the most valuable information in determining the top 10. Again, I am most interested in the search for the top ten, especially since the remaining items are among themselves or even about how the top ten takes place among them.

+5
source share
2 answers

, , ( ). (O (n ^ 2)) , , k- . , ( ) k . : http://en.wikipedia.org/wiki/Double-elimination_tournament.

, , .

+1

Go Chess. MacMahon Algorithms, . , 10 .

, ( , ).

+1

More articles:


All Articles