Project based on the paper Deep Neural Network Ensembles for Time Series Classification (2019) by Hassan Ismail Fawaz, Germain Forestier, Jonathan Weber, Lhassane Idoumghar, Pierre-Alain Muller
The project uses Jupyter notebooks which combine the python sources.
File | Description |
---|---|
Ensemble.py | Provides the Ensemble class which allows combining trained models into an ensemble |
Ensemblebuilder.py | Provides methods to evaluate multiple Ensembles |
Evaluation.py | Provides helper methods to display results |
Helpers.py | Provides useful methods |
LoadData.py | Provides methods to load the datasets |
ModelBuilder.py | Provides methods that create keras models |
PreprocessData.py | Provides means to add noise to datasets |
TrainProdecure.py | Provides methods to train models |
Install requirements and run the respective Jupyter notebook