Purpose:

Compute BoTW feature relevance scores for the AVEC'17 challenge package (easily extendable to AVEC'18 challenge).

How to:

1. Copy-paste InvestigateBoTW.py to <path-to-AVEC-package>/AVEC2017/emotion_baseline_scripts/.

2. Open up your favourite python IDE or a terminal.

3. python InvestigateBoTW.py

Dependencies:

pandas, numpy, matplotlib, keras, innvestigate

Citation:

If you use this code, please star the repository and cite the following paper:

Pandit, Vedhas, Maximilian Schmitt, Nicholas Cummins, and Björn Schuller. "I know how you feel now, and here's why!: demystifying time-continuous high resolution text-based affect predictions in the wild" In 2019 IEEE 32nd International Symposium on Computer-Based Medical Systems (CBMS), pp. 465-470. IEEE, 2019.

@inproceedings{pandit2019know,
  title={{I know how you feel now, and here's why!: Demystifying Time-continuous High Resolution Text-based Affect Predictions In the Wild}},
  author={Pandit, Vedhas and Schmitt, Maximilian and Cummins, Nicholas and Schuller, Bj{\"o}rn},
  booktitle={32nd International Symposium on Computer-Based Medical Systems (CBMS)},
  pages={465--470},
  year={2019},
  organization={IEEE}
}

Additional Notes:

The code posted here should work fine. If you need some advice, tips, documentation, or just interested in building on top of this work in any way, please feel free to reach out.

We are happy to help and/or collaborate. Stalk us to find our postal address to send us a snail-mail, or wait... Better yet, let's save you all that trouble! Here're my email addresses for this project. Happy coding! :)

Contact:

Vedhas Pandit ([email protected], [email protected])

Chair: https://www.informatik.uni-augsburg.de/lehrstuehle/eihw/