search

stripe-archive / brushfire

Distributed decision tree ensemble learning in Scala
Other
392 stars 50 forks source link

make the feature name type generic #1

Closed avibryant closed 9 years ago

avibryant commented 9 years ago

My main motivation here is that in the future, this should allow us to have some kind of enumerated representation of feature names which serializes compactly; once we have that, it probably makes sense to optionally/always forceToDisk the trainingData to get a compact representation we can scan over N times when expanding.

cc @snoble @colinmarc @danielhfrank

snoble commented 9 years ago

"What did you accomplish yesterday?" "I added a semicolon" "What did you accomplish today?" "I removed a semicolon"