Open EmilHvitfeldt opened 2 years ago
I see two implementation of this.
step_celery()
step_kmeans()
step_hclust()
The first option is nice because we don't add that many steps. The second option is nice because it would allow for hyperparameter tuning
step_kmeans is already implemented in the MachineShop package. step_hclust would be nice.
I see two implementation of this.
step_celery()
which takes cluster specificationstep_kmeans()
,step_hclust()
etc etcThe first option is nice because we don't add that many steps. The second option is nice because it would allow for hyperparameter tuning