A Deep Neural Network Method for Heart Rate Variability-Based Multiclass Stress Detection

Abstract

Expectations can naturally cause stress, particularly if such expectations are perceived as hazardous or damaging. Chronic, long-term stress raises the likelihood of mental health conditions like sleeplessness, depression, and anxiety. A popular stress metric is heart rate variability (HRV), which shows changes in the intervals between heartbeats as opposed to heart rate, which is an average. This paper investigates heart rate variability (HRV) as a stress biomarker and suggests a convolutional neural network (CNN)-based model for multi-class stress classification in order to distinguish between no stress, interruption stress, and time pressure stress. The model outperformed current methods in terms of accuracy when tested on the SWELL-KW dataset. This work highlights the significance of HRV properties for stress diagnosis using variance analysis.