Eval importance only on held out data

[falkben]

[MrAE]

@jovo Does this mean we want to compute the importance using only the Out-Of-Bag data during training? Or are we wanting to be able to compute feature importance on new data using a pre-trained forest?

[jovo]

OOB.

On Mon, Jan 21, 2019 at 9:55 AM Jesse Leigh Patsolic < [email protected]> wrote:

@jovo https://github.com/jovo Does this mean we want to compute the importance using only the Out-Of-Bag data during training? Or are we wanting to be able to compute feature importance on new data using a pre-trained forest?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/neurodata/R-RerF/issues/93#issuecomment-456100892, or mute the thread https://github.com/notifications/unsubscribe-auth/AACjchLElKfCYyNHv64ZZVXKu9oJ5zj1ks5vFdTogaJpZM4YpZd0 .

-- the glass is all full: half water, half air. neurodata.io