search

jasonlaska / spherecluster

Clustering routines for the unit sphere
https://medium.com/@jaska_at_clara/simple-datetime-disambiguation-fd2374ce664a
MIT License
331 stars 78 forks source link

Add `normalize=True` parameter #1

Closed jasonlaska closed 7 years ago

jasonlaska commented 8 years ago

Add normalize=True parameter that normalizes data (optional to user) to both classes so that check_estimator can be applied in tests.

brandonrobertz commented 7 years ago

Are there any plans to do this? I assume you'd use l2 norm on the input data (divide vector by its l2 norm)? I assumed spherecluster already did this and got bit. Will be implementing this soon myself.

jasonlaska commented 7 years ago

Until this is implemented, you should be able to use sklearn's built-in normalize() to normalize your vectors (or the equivalent class version which can be included easily in an sklearn pipeline)

Cheers Jason

On Thu, Jun 29, 2017 at 5:35 PM B Roberts notifications@github.com wrote:

Are there any plans to do this? I assume you'd use l2 norm on the input data? I assumed spherecluster already did this and got bit. Will be implementing this soon myself.

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/clara-labs/spherecluster/issues/1#issuecomment-312115431, or mute the thread https://github.com/notifications/unsubscribe-auth/ABk3vIjwmUwH99FZS_DhqLHW6k88VoG4ks5sJBi8gaJpZM4Jhs3G .

-- Sent from Gmail Mobile

brandonrobertz commented 7 years ago

Cool. Thanks, that's what I'll do then.

jasonlaska commented 7 years ago

Addressed by #5, pull version 0.1.5 or higher for this feature.