Hacker News Reader

Thanks :)

I am working on an intelligent feed reader which keeps track of what kind of articles you like and show articles which are of your interest.

To learn user's interest, he has to like/dislike an article.

View on HN

I've long considered using bayesian classification to mark feeds for like/dislike so the system can learn what I like, but from what I hear, the amount of text in the feed may not be sufficient to classify properly. I have previously looked at http://uclassify.com that provides such a web-service that classifies text based on various criteria. I also remember signing up for an actual rss reader web-service that planned to use bayesian classification to learn your likings a couple of years, but they never released -- I don't recall the site name right now.

Is this a personal project or something for JackBe (I notice you work there).

Be mindful that many have tried to build this kind of system and there are a lot of pitfuls outside of just the technology - such as explicit ranking versis implicit/gestures (although I notice you are a UX designer, so you may be thinking of interface issues already).

I've got some experience in this area if you are interesting in connecting - my email is in my profile.

Reply on news.ycombinator.com