Model Ensembles

Is it possible to use KNIME to generate model ensembles?

Let's say for example I use Decision Tree, K nearest Neighbour and Neural Network models and want to combine their predictions into one so each model gets to "vote" on how to classify each case.

Is that possible in KNIME?




Hi Paul,

No, there is not node in KNIME that automatically combines those classification outputs. Some of our learner, i.e. Decision Tree, MLP etc., output the class probabilities for all possible values (in the target column) those can be joined and later combined into a single target column. Maybe this workaround helps to solve your problem?

Regards, Thomas