datatable.models.Ftrl.negative_class¶
negative_class
negative_class
An option to indicate if a “negative” class should be created
in the case of multinomial classification. For the “negative”
class the model will train on all the negatives, and if
a new label is encountered in the target column, its
weights are initialized to the current “negative” class weights.
If negative_class
is set to False
, the initial weights
become zeros.
This option is read-only for a trained model.
Parameters¶
return
bool
Current negative_class
value.
new_negative_class
bool
New negative_class
value.
except
ValueError
The exception is raised when trying to change this option for a model that has already been trained.
The content on this page is licensed under the Creative Commons Attribution 4.0 License
(CC BY 4.0) ,
and code samples are licensed under the MIT License.