Multi-Class Stress Detection Through Heart Rate Variability Using Deep Learning

D. Jayasoniya , KOTHAMASU DIVYA VENKATA SAI LAKSHMISNEHA , METIKALA JYOTHSNA , JAHNAVI SAI MUTHYALA , KADAVAKUDURU SWATHI

Heart Rate Variability (HRV); Stress Detection; 1D CNN; Deep Learning; SWELL-KW Dataset; ANOVA Feature Selection; Time-Domain Features; Frequency-Domain Features; Multi-Class Classification.

Stress, a prevalent psychological and physiological condition, significantly impacts mental well-being and physical health. Detecting stress accurately in its early stages can help prevent chronic disorders and improve quality of life. This study presents a deep learning-based approach for multi-class stress detection using Heart Rate Variability (HRV) data. By leveraging the SWELL-KW dataset, a 1D Convolutional Neural Network (CNN) is implemented to classify stress levels into three categories: No Stress, Interruption Stress, and Time Pressure Stress. The model integrates robust preprocessing, ANOVA-based feature selection, and optimized hyperparameters to enhance performance. Experimental results demonstrate a classification accuracy of 99.9%, significantly outperforming traditional models like SVM and Random Forest. The proposed system shows promise in providing a real-time, non-invasive solution for stress monitoring, thereby contributing to advancements in mental health technology and workplace wellness.

Volume 10 Issue 9, September-2025

Pages : a108-a112