I'm trying to build a content-control (web filtering) application using machine learning (just for training purposes). For example define gaming sites. I'm somewhat familiar with machine learning algorithms, but I have almost no idea how to get features from HTML pages. Can anyone suggest any good material? Maybe some resource with training data? I will be glad of any advice!
You can try WEKA and histogram analysis of your data.
Then, you can use a algorithm for getting the most frequent terms.