It would be useful to have a python script that gathers up all models in a folder and uses them to build an ensemble. This should be possible by taking the results of the training data from the "raw" models and using that as training data for the ensemble. But it will require some infrastructure to make this as robust as possible.
It would be useful to have a python script that gathers up all models in a folder and uses them to build an ensemble. This should be possible by taking the results of the training data from the "raw" models and using that as training data for the ensemble. But it will require some infrastructure to make this as robust as possible.
http://mlwave.com/kaggle-ensembling-guide/