Bug 102629 - highlight interesting articles with bayes filter
Summary: highlight interesting articles with bayes filter
Status: CONFIRMED
Alias: None
Product: akregator
Classification: Applications
Component: general (show other bugs)
Version: unspecified
Platform: Debian testing Linux
: NOR wishlist
Target Milestone: ---
Assignee: kdepim bugs
URL:
Keywords:
: 163627 (view as bug list)
Depends on:
Blocks:
 
Reported: 2005-03-28 01:41 UTC by Fred Schättgen
Modified: 2021-03-09 04:11 UTC (History)
4 users (show)

See Also:
Latest Commit:
Version Fixed In:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Fred Schättgen 2005-03-28 01:41:48 UTC
Version:            (using KDE KDE 3.4.0)
Installed from:    Debian testing/unstable Packages

I'm a bit undecided if I want akregator to pop up a notification window for new articles or not. So it would be cool if akregator could classify the articles with a bayes filter into interesting and less interesting stuff.
Then I could tell akregator to bug me only when there are real news coming in and let it highlight the interesting stories.

I don't know how difficult it is to implement a Bayes filter, but akregator could probably reuse the spam classification tools which kmail is using (spambayes etc.).

The articles then have to be rated somehow. Therefore it would be good to also have two rate-and-advance-to-next-unread shortcuts to make it as convenient as possible to rate an article.
Comment 1 Frank Osterfeld 2005-05-12 23:00:40 UTC
That'd be an interesting feature, indeed. I was also thinking about this. Combined with the Tenor framework and context identification ("what does the user right now?") it could even filter for articles interesting in the current context =) Well, that's more the vision than a concrete plan. 
But simple filtering could be added when we have search folders.

I wonder how to train the filter. It could be trained somehow by observing what kind of articles the user reads and the filter for similar articles, or let the user train it by rating articles. But then rating has to be really simple, otherwise nobody will do it. (maybe rate it by pressing "0" - "9" keys?)
Comment 2 Frank Osterfeld 2005-07-03 00:02:40 UTC
Dennis plans to implement something like this: see http://akregator.pwsp.net/blog/?p=17 for details
Comment 3 Aldo "xoen" Giambelluca 2009-03-07 02:00:44 UTC
Great idea, it would be fantastic. It would be a big step forward usability and time saving for all. 

I think the 1-10 or 5 stars rating it's not the best, I would like to read articles as always and click on a "Interesting" button if I like it, of course I need also a "Not Interesting" button to mark an article as not interesting (for example because akregator marks an article as interesting but it is not).
In this way the user can teach akregator which articles are interesting putting into a virtual folder (Bug 85615, depends on it? probably yes) , wow!
Comment 4 Christophe Marin 2009-10-07 17:58:05 UTC
*** Bug 163627 has been marked as a duplicate of this bug. ***
Comment 5 Justin Zobel 2021-03-09 04:11:21 UTC
Thank you for the bug report.

As this report hasn't seen any changes in 5 years or more, we ask if you can please confirm that the issue still persists.

If this bug is no longer persisting or relevant please change the status to resolved.