Deep Neural Network Ensembles for Time Series Classification

This is the companion repository for our paper also available on ArXiv titled "Deep Neural Network Ensembles for Time Series Classification". This paper has been accepted at the IEEE International Joint Conference on Neural Networks (IJCNN) 2019.

Approach

Data

The data used in this project comes from the UCR/UEA archive, which contains the 85 univariate time series datasets.

Code

The code is divided as follows:

• The main.py python file contains the necessary code to run all experiements.
• The utils folder contains the necessary functions to read the datasets and manipulate the data.
• The classifiers folder contains eight python files one for each deep individual/ensemble classifier presented in our paper.

To run a model on all datasets you should issue the following command:

python3 main.py

To control which datasets and which individual/ensemble classifiers to run see the options in constants.py.

You can control which algorithms to include in the ensemble by changing this line of code.

Prerequisites

All python packages needed are listed in pip-requirements.txt file and can be installed simply using the pip command.

• numpy
• pandas
• sklearn
• scipy
• matplotlib
• tensorflow-gpu
• keras
• h5py
• keras_contrib

Results

The following table shows the results of four ensembles, the raw results can be found here.

Critical difference diagrams

If you would like to generate these diagrams, take a look at this code!

Reference

If you re-use this work, please cite:

@InProceedings{IsmailFawaz2019deep,
  Title                    = {Deep Neural Network Ensembles for Time Series Classification},
  Author                   = {Ismail Fawaz, Hassan and Forestier, Germain and Weber, Jonathan and Idoumghar, Lhassane and Muller, Pierre-Alain},
  booktitle                = {IEEE International Joint Conference on Neural Networks},
  Year                     = {2019}
}

Acknowledgement

We would like to thank NVIDIA Corporation for the Quadro P6000 grant and the Mésocentre of Strasbourg for providing access to the cluster.