« The value of "pretty" | Main | Browser "embed" behavior »
February 19, 2007
MyFeedz clustering
MyFeedz clustering: This weekend Adobe Labs launched an online news-reading service called MyFeedz. I had tried it before but didn't really understand it, and it wasn't until this weekend that I really dug into it. It's not really "an RSS newsreader" so much as it is "A Social Newspaper" -- its focus is on surfacing articles of particular interest to you, rather than just presenting all the stuff that selected others write. I've already found many relevant articles that I would not have otherwise found -- that's valuable to me. MyFeedz feels like a combination of trusted-sources reader (you can add feeds that you'd like to be valued highly), and river-of-news (multiple authors are aggregated together into a single topic's display), and live search engine ("tags" seem often based on text analysis in addition to blog categories and member tagging). Unlike other blogsearch engines these days, the number of false positives is very low, with few splogs or low-content posts. I'm getting a high number of off-topic posts at the moment ("apollo" entries often talk about astronauts, eg), but the learning and personalization cues may minimize this with time spent voting. I'm still not sure how much of the blogosphere the engine samples, nor how individual voting or total voting will cause the system to adapt, but it looks like a valuable addition to my own daily workflow. Check it out and drop a note to the creators at the Labs Forum, if you get the chance... I think there's something here.
Posted by JohnDowdell at February 19, 2007 09:51 PM
Trackback Pings
TrackBack URL for this entry:
http://weblogs.macromedia.com/mtadmin/mt-tb.cgi/8396
Comments
John,
Please compare the machine extracted tags at www.wizag.com with those at myfeedz. We are able to extract and rank more informative tags using language processing and attention data.
Posted by: Ping Liang at April 16, 2007 09:15 PM