Simple form of Machine Learning/AI with Perl?

Cody Fendant has asked for the wisdom of the Perl Monks concerning the following question:

I would like to attempt to set up a simple type of machine learning system with Perl.

I would like to get a bunch of stories from news RSS feeds and upvote and downvote each one as I read them, like on Reddit, and have the system "learn" what kind of stories I like to read.

Any kind of pointers gratefully received.

Comment on Simple form of Machine Learning/AI with Perl?

Replies are listed 'Best First'.
Re: Simple form of Machine Learning/AI with Perl? (tf-idf) by LanX (Saint) on Mar 04, 2017 at 04:58 UTC
> Any kind of pointers gratefully received. Extract relevant "keywords" for each document with the term frequency–inverse document frequency method (see tf-idf modules)... ... and weight them according to your "likes". New stories with high combined weight of the extracted vector of most relevant keywords are probably those you like. You will need to continue voting while you get them to make the system learn. Cheers Rolf _{(addicted to the Perl Programming Language and ☆☆☆☆ :) Je suis Charlie!} PS: Yes this approach could be used for the monastery too.	[reply]
Re^2: Simple form of Machine Learning/AI with Perl? (tf-idf) by Cody Fendant (Hermit) on Mar 04, 2017 at 06:19 UTC
Thanks, that's a promising start.	[reply]
Re: Simple form of Machine Learning/AI with Perl? by 1nickt (Canon) on Mar 04, 2017 at 09:49 UTC
See Machine Learning on Perl, new last week, bit more in depth than your description, maybe. Hope this helps! The way forward always starts with a minimal test.	[reply]