eiron wrote:
] It seems like a nice idea, but it doesn't this approach
] probably wouldn't work well with blogs. Spam may be easily
] classified by bayesian filters. The content of two blog
] entries, however, could easily contain many common keywords,
] yet provide significantly different levels of interest to the
] reader.

I agree... However, thanks for recommending the article. These perl libraries are useful for something else: Figuring out when articles on MemeStreams are associated with the same subject...

RE: XML.com: Working with Bayesian Categorizers [Nov. 19, 2003]