tl;dr: build a randomized decision forest; instances that often end up alone in leaves high in the tree (ie after few splits) are anomalies.
-
-
-
It seems like a similar average-terminal-depth metric could be an interesting feature importance metric for standard random forests.
-
@avibryant wonder if a metric of connection density might do the same for neural nets
End of conversation
New conversation -
-
-
@avibryant Very neat. I'm curious where you found this paper? -
@armon from this tweet about it getting added to scikit:https://twitter.com/glouppe/status/657978716790870016 … -
@avibryant Cool, thanks!
End of conversation
New conversation -
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.