tensorflow / decision-forests

A collection of state-of-the-art algorithms for the training, serving and interpretation of Decision Forest models in Keras.
Apache License 2.0
663 stars 110 forks source link

does it support multi_labels classfication? #19

Closed Bigmai-1234 closed 3 years ago

achoum commented 3 years ago

No. Multi-label classification is not supported (i.e. training a model that predicts several independent labels).

The recommended solution is to train an individual TF-DF model for each label (possibly in parallel).

Remark that Decision Forests do not have the capability to extract hidden representations like Neural Networks do. In other words, "multi-head training" is equivalent to independent training.

Ps: Multi-class classification is supported (i.e. training a classification model on a label with more than two possible classes).

Bigmai-1234 commented 3 years ago

got it, thanks!@ach

No. Multi-label classification is not supported (i.e. training a model that predicts several independent labels).

The recommended solution is to train an individual TF-DF model for each label (possibly in parallel).

Remark that Decision Forests do not have the capability to extract hidden representations like Neural Networks do. In other words, "multi-head training" is equivalent to independent training.

Ps: Multi-class classification is supported (i.e. training a classification model on a label with more than two possible classes).

got it! thanks.