Comments on: An algo for Twitter Authority http://mayank.name/2008/12/27/an-algo-for-twitter-authority/ Thoughts on Business, Marketing and Social Media Sun, 07 Mar 2010 04:31:46 +0000 http://wordpress.org/?v=2.9.2 hourly 1 By: TwitterRank: A rank based algorithm for detecting authority in twitter « Informatio http://mayank.name/2008/12/27/an-algo-for-twitter-authority/comment-page-1/#comment-537 TwitterRank: A rank based algorithm for detecting authority in twitter « Informatio Mon, 08 Jun 2009 06:02:30 +0000 http://mayank.name/blog/?p=294#comment-537 [...] Mayank proposes his algorithm: http://mayank.name/2008/12/27/an-algo-for-twitter-authority/ [...] [...] Mayank proposes his algorithm: http://mayank.name/2008/12/27/an-algo-for-twitter-authority/ [...]

]]> By: Denis Florent http://mayank.name/2008/12/27/an-algo-for-twitter-authority/comment-page-1/#comment-93 Denis Florent Fri, 20 Mar 2009 16:50:41 +0000 http://mayank.name/blog/?p=294#comment-93 I chose different options, and came up with this test : http://www.denisflorent.fr/twitter-authority-algorithm-proposal I chose different options, and came up with this test :

http://www.denisflorent.fr/twitter-authority-algorithm-proposal

]]> By: Qiu Foreign Exchange http://mayank.name/2008/12/27/an-algo-for-twitter-authority/comment-page-1/#comment-97 Qiu Foreign Exchange Wed, 21 Jan 2009 13:24:02 +0000 http://mayank.name/blog/?p=294#comment-97 the (followers/following) ratio is not correct, since in this case, the one with 100 followers and 100 following gets the same score as the one with 1 followers and 1 following. the (followers/following) ratio is not correct, since in this case, the one with 100 followers and 100 following gets the same score as the one with 1 followers and 1 following.

]]> By: mayank http://mayank.name/2008/12/27/an-algo-for-twitter-authority/comment-page-1/#comment-96 mayank Sun, 28 Dec 2008 14:23:13 +0000 http://mayank.name/blog/?p=294#comment-96 @Cail: Agreed but in Loic's post he gave LeWeb's example where about 7000 tweets were made and his problem was that he wanted to only skim through tweets of people who have higher authority. For a case like LeWeb search by authority of a person is relevant and otherwise search by authority of a tweet. @funkyboy: Nice post and yes, people always try to game such things :) @Cail: Agreed but in Loic’s post he gave LeWeb’s example where about 7000 tweets were made and his problem was that he wanted to only skim through tweets of people who have higher authority. For a case like LeWeb search by authority of a person is relevant and otherwise search by authority of a tweet.

@funkyboy: Nice post and yes, people always try to game such things :)

]]> By: funkyboy http://mayank.name/2008/12/27/an-algo-for-twitter-authority/comment-page-1/#comment-94 funkyboy Sun, 28 Dec 2008 12:58:22 +0000 http://mayank.name/blog/?p=294#comment-94 I wrote down a set of heuristics. Maybe they can be helpful to calculate the authority. http://www.louisgray.com/live/2008/12/posty-authors-guide-to-unfollow.html I think the algorithm should be secret, otherwise people will try to cheat :) I wrote down a set of heuristics.
Maybe they can be helpful to calculate the authority.

http://www.louisgray.com/live/2008/12/posty-authors-guide-to-unfollow.html

I think the algorithm should be secret, otherwise people will try to cheat :)

]]> By: Cail http://mayank.name/2008/12/27/an-algo-for-twitter-authority/comment-page-1/#comment-95 Cail Sat, 27 Dec 2008 13:46:50 +0000 http://mayank.name/blog/?p=294#comment-95 What you do is trying to calculate the authority of a person. The advantage of blogging and microblogging however is that authority could be assigned to individual posts and tweets. So don't take the *total* numbers of RTs in relation to the *total* number of tweets, but rather the number of RTs per tweet in relation to the number of followers. Etc. (needs some refinement in order to prevent those with 1/1 being the top-listed tweets.. so to some extent, the total number of followers should be weighed in, too) What you do is trying to calculate the authority of a person. The advantage of blogging and microblogging however is that authority could be assigned to individual posts and tweets.

So don’t take the *total* numbers of RTs in relation to the *total* number of tweets, but rather the number of RTs per tweet in relation to the number of followers. Etc. (needs some refinement in order to prevent those with 1/1 being the top-listed tweets.. so to some extent, the total number of followers should be weighed in, too)

]]>