This repository contains the code to reproduce the ML analysis in
@Article{electronics10050613,
AUTHOR = {Tervonen, Jaakko and Pettersson, Kati and Mäntyjärvi, Jani},
TITLE = {Ultra-Short Window Length and Feature Importance Analysis for Cognitive Load Detection from Wearable Sensors},
JOURNAL = {Electronics},
VOLUME = {10},
YEAR = {2021},
NUMBER = {5},
ARTICLE-NUMBER = {613},
URL = {https://www.mdpi.com/2079-9292/10/5/613},
ISSN = {2079-9292},
DOI = {10.3390/electronics10050613}
}
Please cite the paper above if you use any part of this code in a publication.
The code was developed on Xubuntu 20.04 LTS and Python 3.6.9. The required packages are listed in requirements.txt, you can install them with pip install -r requirements.txt.
The dataset used is the CogLoad dataset available here. To use default paths, place the data of each individual user under ./data/train/
Usage:
To extract features run python data.py and to run Bayesian hyperparameter optimization for XGBoost model, run python xgb_hyperopt.py.