To try out the classifier on real data and show the different ways it can be used, you can apply it to entries from a blog or other RSS feed. To do this, you'll need to get the Universal Feed Parser, which we used in Chapter 3. If you haven't already downloaded it, you can get it from http://feedparser.org. More information on installing the Feed Parser is given in Appendix A.
Although a blog will not necessarily contain spam in its entries, many blogs contain some articles that interest you and some that don't. This can be because you only want to read articles in a certain category or by a certain writer, but it's often more complicated than that. Again, you can set up specific rules for things that do and do not interest you—maybe you read a gadget blog and are not interested in entries that contain the word "cell phone"—but it's much less work to use the classifier you've built to figure out these rules for you.
A benefit of classifying entries in an RSS feed is that if you use a blog-searching tool like Google Blog Search, you can set up the results of your searches in a feed reader. Many people do this to track products, things that interest them, even their own names. You'll find, though, that spam-based or useless blogs trying to make money from blog traffic can also appear in these searches.
For this example, you can use any feed you like, although many feeds have too few entries to do any effective training. This particular example uses the results of ...