Traffic4Cast2020-Deep-Sedenion-Network

This resipository contains our code submitted to Traffic4cast2020 competition (https://www.iarai.ac.at/traffic4cast/2020-competition/challenge/)

To generate our submitted test inference, run test_sedenion.py To retrain the model, run sedenion_trainer.py

This work is made available under the attached license

If using this code or work presented in the paper https://arxiv.org/abs/2012.03874 please cite

@article{bojesomo2020traffic,
      title={Traffic flow prediction using Deep Sedenion Networks}, 
      author={Alabi Bojesomo and Hasan Al-Marzouqi and Panos Liatsis },
      year={2020},
      eprint={2012.03874},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}

@article{bojesomo2024hypercomplex,
  author={Bojesomo, Alabi and Liatsis, Panos and Marzouqi, Hasan Al},
  journal={IEEE Signal Processing Magazine}, 
  title={Deep Hypercomplex Networks for Spatiotemporal Data Processing: Parameter efficiency and superior performance [Hypercomplex Signal and Image Processing]}, 
  year={2024},
  volume={41},
  number={3},
  pages={101-112},
  keywords={Training;Convolution;Algebra;Quaternions;Image processing;Data processing;Spatiotemporal phenomena},
  doi={10.1109/MSP.2024.3381808}}