May 09, 2023
Gary Robinson
agent
wrote
LityxIQ may still do binning in some situations depending on other settings like choosing numeric missing replaced with categories or having too many unique values in a field.
It will bin for efficiency sake if the field has too many categories. Specifically, some/most algorithms do not scale well if the number is more than 25 to 30. For numerics, if "categorize" is checked in the missing variable area, then it will be binned if there are missing values in the field, even if the algorithm settings didn't want binning.
So by choosing to categorize numeric missing it essentially turns on numeric binning for those numeric fields with a missing value. For numeric binning, there is no other way to do it. You can’t have some values categorized and others stay as numbers. And you also can’t just leave missing values. They have to be removed or imputed if you don’t want things categorized. But if there are no missing values in the field, it won’t be categorized with that setting.
It will bin for efficiency sake if the field has too many categories. Specifically, some/most algorithms do not scale well if the number is more than 25 to 30. For numerics, if "categorize" is checked in the missing variable area, then it will be binned if there are missing values in the field, even if the algorithm settings didn't want binning.
So by choosing to categorize numeric missing it essentially turns on numeric binning for those numeric fields with a missing value. For numeric binning, there is no other way to do it. You can’t have some values categorized and others stay as numbers. And you also can’t just leave missing values. They have to be removed or imputed if you don’t want things categorized. But if there are no missing values in the field, it won’t be categorized with that setting.
No answers