Electrodermal Activity for Stress Identification

Dataset Description

The "Electrodermal Activity for Stress Identification" dataset is associated with the study "Deep Support Vector Machines for the Identification of Stress Condition from Electrodermal Activity" and provides physiological data used to classify stress and calm states based on electrodermal activity (EDA) signals.

Dataset Details

Source: The data was collected from 147 volunteers who participated in an experiment designed to induce stress and calm conditions through audiovisual stimuli exposure.

Recording Device: The Empatica E4 wristband was used, a commercially available and validated device for electrodermal activity measurement.

Included Variables:

Raw EDA signals: Continuous recordings of electrodermal activity measured in microsiemens (µS).

Extracted Features: Processed data in the time domain, frequency domain, and morphological analysis.

State Labels: Each data segment is classified as stress or calm, based on the experimental condition.

Time Stamps: Time markers for each recording in seconds, tracking the signal evolution during stimulus exposure.

File Format: The dataset is available in CSV and MAT formats, making it compatible with Python, MATLAB, and other data science tools.

Usage and Applications

This dataset is useful for research in computational neuroscience, machine learning, psychophysiology, and digital health. It can be applied to:

Training and validating stress classification models using machine learning algorithms such as SVM, neural networks, and deep learning.

Analyzing physiological stress patterns in controlled environments.

Developing biofeedback systems and mental well-being applications for real-time stress detection.

Open Access and Availability

The dataset has been published in open access on the Zenodo repository, allowing for reuse and replicability in future research. 

License

This dataset is available under a Creative Commons (CC-BY 4.0) license, permitting usage, modification, and redistribution as long as proper citation is given to the original source.