Ensemble Tree methods

Ensemble methods combine several Decision Trees in order to achieve a better prediction performance. The basic principle is to train not just one model (one tree), but several trees – a forest of trees – and to combine the individual results into a common result.

There are basically two technologies with which an ensemble of trees can be created.

Bagging

The bagging methods include:

• Random Forest
• Extra Trees

Boosting

The boosting methods include:

• Gradient Boosting
• Histogram Gradient Boosting
• XGBoost
• LightGBM

Unsupported additional Ensemble Tree methods The models BaggingClassifier, BaggingRegressor, AdaBoostClassifier and AdaBoostRegressor are also available in Scikit-learn. During an ONNX export they currently generate a graph that is incompatible with TwinCAT libraries, which means they cannot be supported.