tag:blogger.com,1999:blog-4999686744433629981.post2464652489783715596..comments2024-02-14T22:06:08.851-08:00Comments on Augmented Social Cognition Research Blog from PARC: Short and Tweet: Experiments on Recommending Content from Information Streams (Specifically, Twitter)Ed H. Chihttp://www.blogger.com/profile/06360447323238002978noreply@blogger.comBlogger7125tag:blogger.com,1999:blog-4999686744433629981.post-35058627334852011492010-05-03T22:26:57.002-07:002010-05-03T22:26:57.002-07:00This comment has been removed by a blog administrator.Amitnoreply@blogger.comtag:blogger.com,1999:blog-4999686744433629981.post-48357031371677505802010-04-29T09:54:37.755-07:002010-04-29T09:54:37.755-07:00@Joe: Thanks for the pointer to EdgeRank. I hadn...@Joe: Thanks for the pointer to EdgeRank. I hadn't seen it yet when you posted it, so it was indeed very useful for comparing our techniques to it. Nice find!Ed H. Chihttps://www.blogger.com/profile/06360447323238002978noreply@blogger.comtag:blogger.com,1999:blog-4999686744433629981.post-63691638263329620052010-04-28T21:56:39.463-07:002010-04-28T21:56:39.463-07:00This comment has been removed by a blog administrator.Amitnoreply@blogger.comtag:blogger.com,1999:blog-4999686744433629981.post-87464120482490001292010-04-23T14:15:26.201-07:002010-04-23T14:15:26.201-07:00FWIW, TechCrunch just posted some details about Ed...FWIW, TechCrunch just posted some details about <a href="http://techcrunch.com/2010/04/22/facebook-edgerank" rel="nofollow">EdgeRank: The Secret Sauce That Makes Facebook’s News Feed Tick</a>.<br /><br />May be of potential interest / relevance to this work.Joe McCarthyhttp://gumption.typepad.comnoreply@blogger.comtag:blogger.com,1999:blog-4999686744433629981.post-64818133879146411492010-04-22T23:50:15.894-07:002010-04-22T23:50:15.894-07:00@Joe: Thanks for the comments. The participation ...@Joe: Thanks for the comments. The participation architecture for Twitter indeed seems to need further study. It could be that we ought to have taken the distribution into account better in our recommender design.<br /><br />The use of 'followee' was simply one of practicality. We kept getting confused ourselves when we talked to each other, so we had to adopt better terminology during the research!<br /><br />@Drivelry: Indeed, thinking about Twitter as a way to construct a personalized newspaper is very close to how we thought about our research problem!Ed H. Chihttps://www.blogger.com/profile/06360447323238002978noreply@blogger.comtag:blogger.com,1999:blog-4999686744433629981.post-85643087365149281552010-04-22T19:26:03.931-07:002010-04-22T19:26:03.931-07:00Interesting. This issue is becoming more and more ...Interesting. This issue is becoming more and more important with the breakdown of old 'bundled' news models.<br /><br />In a less structured way I have been recently playing around with <a href="http://www.drivelry.com/finding-the-best-blogs-on-the-web-for-your-personalized-newspaper/588/" rel="nofollow">article recommendation engines / methodologies</a> and also found that Twitter seemed to be the logical source data to build a better recommendation mousetrap from. <br /><br />Somewhat appropriately I also found this very blog (which I've now incorporated in my Calibre/Kindle RSS feed) via a Tweet.@Drivelryhttp://www.drivelry.comnoreply@blogger.comtag:blogger.com,1999:blog-4999686744433629981.post-84328275615979266942010-04-22T17:49:27.433-07:002010-04-22T17:49:27.433-07:00This is very interesting work! I like the fact tha...This is very interesting work! I like the fact that you've incorporated different models that will accommodate a broad range of Twitter user behavior.<br /><br />When <a href="http://themetricsystem.rjmetrics.com/2010/01/26/new-data-on-twitters-users-and-engagement/" rel="nofollow">RJMetrics reported on New Data on Twitter’s Users and Engagement in January 2010</a>, I was surprised to discover how much the data revealed a power law of participation, e.g.,<br /><br />* A large percentage of Twitter accounts are inactive, with about 25% of accounts having no followers and about 40% of accounts having never sent a single Tweet<br /><br />* About 80% of all Twitter users have tweeted fewer than ten times.<br /><br />* Only about 17% of registered Twitter accounts sent a Tweet in December 2009<br /><br />I also like your discussion of the relevance / serendipity tradeoff (in the paper).<br /><br />Finally, I want to commend you for using "followee" rather than "friend" throughout your description of this work, as it makes it much easier for me to follow [yes, pun intended].Joe McCarthyhttp://gumption.typepad.comnoreply@blogger.com