The idea here is that LightGBM is powerful and if all rows are used, the predictions will likely be perfect, and so no anomalous items will be found. The regression object is named sub_model and is ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results