weaponsjtu / Kaggle_xBle

A Swiss Army Knife for Machine Learning Practice, cross validation, model selection, ensemble selection, stacking

Kaggle_xBle

A Swiss Army Knife for Machine Learning Practice, cross validation, model selection, ensemble selection, stacking. It is built based on scikit-learn package.

gen_fold.py

Cross validation

gen_featu.py

Feature Engineering

model_library.py

Hyper parameter optimization, generate hundreds of models using different model parameters

gen_ensemble.py

Ensemble Selection, forward stepwise greedy ensemble with replacement

gen_stacking.py

Model Stacking, Raw feature + meta feature, XGBoost + AdaBoost

param.py

All parameters

PS: need to modify gen_featu.py when apply a new task

https://github.com/deap/deap