I had to look up "splog" (a
spam blog, which the author uses to promote affiliated Web sites, according to Wikipedia) and double-check that "scrobble", the term that Last.fm uses to report to the service the names of the songs you're listening to.
A spam blog is often composed of parts of other blogs and news articles.
Step 1: For initialization, spam blogs in the spam seed are appended in an initial spam blog list SBlog.
SBlog is composed of blogs equal to or under threshold C, which is the lowest spam blog rate.
Step 5: For each extracted spam blog [b.sub.i] [member of] SBlog, sbrate([b.sub.i]) is recalculated.
Step 6: The final spam blog set is null (FSBlog= [empty set]) because not enough possible spam blogs can be extracted.
The blogs are appended in an initial spam blog list.
Performing linguistic analysis on blogs is plagued by two additional problems: (1) the presence of spam blogs and spam comments and (2) extraneous noncontent including blog rolls, link rolls, advertisements, and sidebars.
Over the past year (Kolari 2007) we have developed techniques to detect spam blogs as they fit the overall architecture (figure 5), arrived at through our discussions with practitioners.
Spam blogs, for example, often form communities whose structural properties are very unlike those of naturally occurring blogs.
Such
spam blogs, or "splogs," typically collect money from advertisers as users click on their links, and some visitors purchase the promoted products.