This repository is for supplementary codes used to explore and analyze the K-EmoPhone Dataset.
- Description of the dataset: https://doi.org/10.1038/s41597-023-02248-2
- K-EmoPhone: A Mobile and Wearable Dataset with In-Situ Emotion, Stress, and Attention Labels. 2023. Soowon Kang, Woohyeok Choi, Cheul Young Park, Narae Cha, Auk Kim, Ahsan Habib Khandoker, Leontios Hadjileontiadis, Heepyung Kim, Yong Jeong, Uichin Lee. Scientific Data, 10, 351.
- Dataset URL: https://doi.org/10.5281/zenodo.7606611
We have run this code under the environment as below:
- OS: Ubuntu 20.04 installed with Windows Subsystem for Linux (WSL)
- This code highly depends on a python multiprocessing library, ray which does not fully support Windows OS.
- CPU: AMD Ryzen 9 5900x 12-Core
- This is not mandatory; you can run this code (with a minor modification) although you have the smaller number of cores.
- RAM: 128GB
- This is not mandatory; we expected about 40GB of RAM to be required (but not tested).
In addition, you need to install conda for managing packages and virtual environment.
-
Download the K-EmoPhone dataset.
-
Download this repository
$ git clone https://github.com/Kaist-ICLab/K-EmoPhone_SupplementaryCodes.git
$ cd K-EmoPhone_SupplementaryCodes-
Replicate our conda environment (environment.yml), referring to this.
-
Run your own Jupyter environment.
-
Then, open analysis.ipynb.