Re:Browser ad-blocking the same way? (Score 1)

I think it could work. Why don't you implement it and let us know how it turns out? The basic implementation is not that complicated. I did something similar for a class project this semester, although I used boosting with decision stumps instead of "Bayesian" learning. The tricky part is selecting the features. There isn't as much text, so you'd also want to look at things like position in the web page, image size, accompanying tags, etc.