I feel like almost all lay discussions regarding classification tasks, which are a very relevant-to-real-life application of stats/ML/AI/whatever-you-call-it, would be improved by people knowing that there is usually a tradeoff involved.
-
-
This is exactly the same problem that exists in medical screening - false positives vs false negatives. Maybe we can learn something from that.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
another good discussion would be: "Hey we've managed to catch 8 bears out of 10 furry animals reliably" "Yes but bears make up 90% of this forest, this system is literally worse than calling everything a bear"
-
if you have a system where fraud happens 0.1% of the time. a 99.9% accurate classification system can be made by classifying everything as "not fraud"
End of conversation
New conversation -
-
-
All I can think of here is the Japanese counting word for bunnies
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.