The repository belongs to paper that investigates the potential of recognising arousal in motor activity. We formulate arousal detection as a statistical problem of separating two sets - motor activity under emotional arousal and motor activity without arousal.
The repository contains scripts and functions for the preprocessing of the data and the machine learning experiments conducted. Our aim is to keep our research reproducable and transparent.
The data are accessible in the Open Science Framework (OSF) via the following link. All the files should be stored in the folder data, so that the provided scripts have the correctly assigned paths.
The projects uses poetry for dependency management. Instructions on how to install poetry can be found on their webpage.
When poetry is successfully installed, it can be used to install the repository.
cd emotion-extraction
poetry init
The data is loaded, structured as objects and pickled with the script:
emotion_extraction/pickle_objects.py
The machine learning experiments can be run by the script:
emotion_extraction/ml_classification.py