It would be fascinating to see which words had most predictive value in each direction.
-
-
-
Definition of "Mattered" is the definition of classifier bias.
End of conversation
New conversation -
-
-
The more I think about metrics for what matters the more it feels like very little news matter.
-
Facts that cause you to see reality in a different way are the most useful news to me; also hearing about new things.
End of conversation
New conversation -
-
-
Matter to whom?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
What definition of "mattered" did you have in mind?
-
Traffic starting a year after publication.
- 1 more reply
New conversation -
-
-
Is this not the premise of a filter bubble?
-
Yeah except the filter is "stories that are important" vs. "stories that make you click"
-
So the feedback is explicit ("This article matters") as opposed to implicit ("This article's title is interesting"). Same concept holds.
-
Identical. Just different loss function
End of conversation
New conversation -
-
-
I think we'd find that most "breaking" news turned out not to matter.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Our feed at
@PressReader learns which stories ppl read most deeply & shows those to similar readers. -
It surfaces v different (hi-qual) content. Not obfuscated by likes/shares & our bias toward sharing click bait or headlines we agree with.
End of conversation
New conversation -
-
-
Isn't it what facebook is doing with everyone's newsfeed?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
You of all people should realize the insanity if you replace "news stories" with "startups." Both are a form of salient events.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
my intuition is that like stock prediction the model would be very prone to overfitting and/or not finding solid correlation between X and y
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Easily replicable with
@monkeylearn An example:https://monkeylearn.com/blog/analyzing-10-years-of-startup-news-with-machine-learning/ …Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Not possible to do with just corpus of published text, requires knowledge of external events and context i.e: sentiment, relationships, etc
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.