4 ms·
...then another problem is treating headlines as if they are all the same. They're not. What you indicated was that headlines are written in different circumst
by misterbowfinger 10y ago
...then another problem is treating headlines as if they are all the same. They're not.
What you indicated was that headlines are written in different circumstances. But it doesn't mean that they're not all, roughly, similar.
I betting there exists a strong correlation between clickbait-y-ness and pageviews.
So,
1. Gather all articles with a lot of pageviews
2. Do some NLP to get the really bad ones, i.e. "Look at what this <noun> did after <predicate>"
3. Remove outliers
And bang, you mark all of the sucky articles. That could be included in a Chrome extension, much like uBlock. Hell, why not just have it in uBlock?