/gretta

Tensor-based SSA for sparse datasets with spatiotemporal information

Primary LanguagePythonMIT LicenseMIT

GRETTA

Generalized REstricted Tensor Timeseries Analysis.

This package is designed to perform multivariate analysis of incomplete timeseries based on the generalization of the restricted SSA method to sparse higher order (3D) data. See an example on the analysis of spatiotemporal humidity data in the Example-1.ipynb jupyter notebook.

Installation

pip install gretta

Requirements

  • numpy
  • scipy
  • pandas
  • numba

Citation

If you use gretta in published research, please cite:

Frolov E, Oseledets I. 2023. Tensor-Based Sequential Learning via Hankel Matrix Representation for Next Item Recommendations. IEEE Access. 2023 Jan 5; 11:6357-71. DOI: 10.1109/ACCESS.2023.3234863. arXiv: 2212.05720.

BibTex entry:

@ARTICLE{Frolov2023,
  author={Frolov, Evgeny and Oseledets, Ivan},
  journal={IEEE Access}, 
  title={Tensor-Based Sequential Learning via Hankel Matrix Representation for Next Item Recommendations}, 
  year={2023},
  volume={11},
  number={},
  pages={6357-6371},
  doi={10.1109/ACCESS.2023.3234863}}

Acknowledgements

Development of this library is supported by the RSCF Grant 22-21-00911.