EnsemblesTutorial

Python notebooks to accompany the paper A Tutorial on Ensembles in Machine Learningwith Python Examples.
The notebooks are as follows:

1. Ensembles-Preliminaries: Code demonstrating the impact of ensemble size and diversity on accuracy.
2. Ensembles-Bagging: Code for bagging and random subspace ensembles.
3. Ensembles-RandomF: Using random forest to generate feature importance scores and OOB estimates of generalisation accuracy.
4. Ensembles-Boosting: A simple AdaBoost example to illustrate the internal workings.
5. Ensembles=GBoost: Gradient Boosting compared with other ensemble methods.
6. Ensembles-Hetero: A heterogeneous ensemble with 7 estimators complared with a bagging ensemble.
7. Ensemble-Stacking: A comparison of a heterogeneous with some stacking alternatives.

ensemble_functions.py: Some helper functions for ensembles.

Data files

The HotelRevHelpfulness and AthleteSelection datasets are made available under the Open Database License: http://opendatacommons.org/licenses/odbl/1.0/. Any rights in individual contents of the database are licensed under the Database Contents License: http://opendatacommons.org/licenses/dbcl/1.0/
The wine dataset is covered by the copyright provisions of the UCI repository.

1. wine.csv: The wine dataset from the UCI repository.
2. HotelRevHelpfulness.csv: The hotel review dataset - see: https://researchrepository.ucd.ie/handle/10197/1894
3. AthleteSelection.csv: A toy dataset with just two predictive features.