RE: "Warning: Blogs Can Be Infectious"

Heh - thanks for the tipoff Graham, just blogged it.
http://jena.hpl.hp.com:3030/blojsom-hp/blog/technologies/blogging/communitie
s/?permalink=C3FB344B1EA27A85C24DD1F05BEB4529.textile&smm=y

I was aware of the work but not that it had made Wired.

Try typing in hpl to the search demo mentioned in the article and you get my
blog!
http://www-idl.hpl.hp.com/cgi-bin/blogs/search_new.cgi?s=hpl

That's about as near to fame as I can manage :)

Steve

> -----Original Message-----
> From: public-esw-request@w3.org 
> [mailto:public-esw-request@w3.org] On Behalf Of Graham Klyne
> Sent: 08 March 2004 13:15
> To: public-esw@w3.org
> Subject: "Warning: Blogs Can Be Infectious"
> 
> 
> 
> The following spotted in ACM's Technews service, at: 
> http://www.acm.org/technews/articles/2004-6/0305f.html#item5
> 
> Is there any contact between this and HP's semantic blogging work?
> 
> I also wonder if it has any implications for FOAF-related 
> applications.
> 
> #g
> --
> 
> # "Warning: Blogs Can Be Infectious"
> Wired News (03/05/04); Asaravala, Amit
> 
> Researchers at Hewlett-Packard Labs used Intelliseek's BlogPulse Web 
> crawler to mine numerous Weblogs, after which they mapped out the 
> connections and topics shared among a large number of sites. Analysis 
> showed that topics would often appear on a small number of relatively 
> obscure blogs a few days before showing up on more popular 
> sites. "There is 
> a lot of speculation that really important people are highly 
> connected, but 
> really, we wonder if the highly connected people just listen to the 
> important people," explains HP Labs researcher Lada Adamic. The team 
> learned that when an idea "infected" at least 10 blogs, 70 
> percent of those 
> blogs failed to supply links back to another blog that 
> previously mentioned 
> the idea, so the researchers devised methods to deduce the 
> point of origin 
> of information by noting textual, link, and infection rate 
> similarities. 
> "What we're finding is that the important people on the Web are not 
> necessarily the people with the most explicit links [back to 
> their sites], 
> but the people who cause epidemics in blog networks," says HP 
> researcher 
> Eytan Adar. The scientists have encapsulated their techniques 
> into the 
> iRank search algorithm, which ranks sites according to how 
> well they inject 
> ideas into the mainstream. Future plans include making iRank 
> resistant to 
> Google-bomb-type attacks, while some of the team's research 
> is accessible 
> online via the Blog Epidemic Analyzer program. The HP Labs 
> research could 
> help sociologists chart the course of knowledge epidemics, 
> which marketers 
> could also exploit to sell their products directly to the 
> most influential 
> members of a group.
> Click Here to View Full Article: 
> http://www.wired.com/news/culture/0,1284,62537> ,00.html
> 
> 
> 
> ------------
> Graham Klyne
> For email:
> http://www.ninebynine.org/#Contact
>

Received on Monday, 8 March 2004 12:01:34 UTC