HanbaekLyu / TDL

Temporal Dictionary Learning

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Temporal Dictionary Learning

Dictionary Learning from joint time serieses using Online Matrix Factorization (flexible nonnegativity constraints on dictioanry and code matrices) Learns dictionary atoms for short time-evolution patterns of multiple entities and uses them to reconstruct the time-series data

References

These codes are based on the following papers

  1. Hanbaek Lyu, Christopher Strohmeier, Deanna Needell, and Georg Menz, “COVID-19 Time Series Prediction by Joint Dictionary Learning and Online NMF” https://arxiv.org/abs/2004.09112

  2. Hanbaek Lyu, Palina Salanevich, Jacob Li, Charlotte Huang, and Deanna Needell "Temporal Dictionary Learning for EEG and Constructing Correlation Tensor" In preperation.

File description

  1. utils/TDL.py : Main file implementing temporal dictionary learning
  2. utils/TDL_plotting.py : Helper functions for plotting
  3. utils/onmf.py : Online Nonnegative Matrix Factorization algorithm (generalization of onmf to the tensor setting by folding/unfolding operation)
  4. covid_dataprocess.py : Preprocessing functions (modify this for your own data type)
  5. TDL-COVID-Test.ipynb : Jupyter notebook example of temporal dictionary learning

Author

  • Hanbaek Lyu - Initial work - Website

License

This project is licensed under the MIT License - see the LICENSE.md file for details

About

Temporal Dictionary Learning

License:MIT License


Languages

Language:Jupyter Notebook 80.5%Language:Python 19.5%