I was wondering how the 'like' filter works..

Is it purely based on matching the tags liked/disliked, or does it get deeper with like bayesian filtering, or full text or…

It would be really awesome if it smart learned as I read articles, and had an ability to tell it I liked/disliked/was indifferent about the article.

Super epic would be going to deep learning levels, so I don’t have to know why/what I liked about the article, but it just figures it out and learns from the fact I liked it.

1 Like

It is a very simple filter. Any tag, author, site, or partial text of title that is trained up gets put in the focus feed. Multiple matches does nothing. Any tag, etc., that is trained down AND does not match anything trained up gets filtered out of the default feed. Everything is always in the all feed.

Thanks for that :slight_smile: Was hoping it was going to be something ‘smarter’/more involved :frowning: