linfa
linfa copied to clipboard
enhancement preprocessing
support OnehotEncoder, LabelEncoder, OrdinalEncoder like sklearn.
I just tried to implement this feature, but the dataset seems to be bound to the float property, which makes it impossible to create datasets (as input) with non-numeric values. Due to the structure of the fit trait, I am unfortunately bound to the input of a dataset. Any help would be appreciated.
What Dataset
methods are you trying to call that are bound by Float
? You can access the records and targets of a Dataset
via the records()
and targets()
methods respectively, even if the data type isn't Float
. Typically those methods are all you need to implement Fit
.