Apply variable-length PSO to find the optimal structure of Weighted Deep Ensemble
- conda create --name py36 python=3.6
- conda install -c anaconda scikit-learn
- conda install -c conda-forge xgboost
- conda install pandas
- Create file
data_helper.py
:
data_folder = 'path/to/data/folder'
file_list = ['abalone', 'balance',...]
- Run:
- VLPSO:
python demo_weighted_gcforest_vlpso.py