Scholarlink

In recent decades, mental health has become a concern due to the prevalence of chronic stress, which can lead to various health problems. One promising non-invasive method that has been widely used to detect stress is electro-dermal activity (EDA). EDA is considered an accurate and sensitive stress measurement method that analyzes skin conductance (SC) due to changes in sympathetic nerve activity. The SC signal consists of phasic and tonic com-ponents, so it is necessary to deconvolve it when analyzing stress conditions. In this study, a convex optimization method (cvxEDA) was proposed to de-compose and analyze EDA signals to assess the emotional state of chronic stress sufferers. This method can help identify different types of stress, such as physical, cognitive, or emotional stress, and aid in handling them according to psychological conditions. Stress-type classification will be performed using a Multi-Layer Perceptron (MLP) algorithm. The study involved 20 subjects who underwent 3 measurement sessions with different stress stimuli to increase their stress levels. The study results demonstrated that cvxEDA method successfully identified and separated the SC components. The phasic component (SCR) was found to describe the subjects’ physiological responses to the stressors. The MLP model was used to classify 3 different stress types and achieved better accuracy value in the classification of cognitive stress with an accuracy of 82% and an F1-score of 80.13% for cognitive stress. This research provides implications in the form of the potential for early detection of stressful situations so that it can provide treatment according to the stress-type.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Classification of Stress Types Based on Electrodermal Activity Signal Using Multi-layer Perceptron Method

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Clara Tyas Ratri,
  • Osmalina Nur Rahma,
  • Sayyidul Istighfar Ittaqillah,
  • Rima Firda Alfisyah,
  • Atin Asna Octavia,
  • Fadli Ama,
  • Khusnul Ain,
  • Rifai Chai

摘要

In recent decades, mental health has become a concern due to the prevalence of chronic stress, which can lead to various health problems. One promising non-invasive method that has been widely used to detect stress is electro-dermal activity (EDA). EDA is considered an accurate and sensitive stress measurement method that analyzes skin conductance (SC) due to changes in sympathetic nerve activity. The SC signal consists of phasic and tonic com-ponents, so it is necessary to deconvolve it when analyzing stress conditions. In this study, a convex optimization method (cvxEDA) was proposed to de-compose and analyze EDA signals to assess the emotional state of chronic stress sufferers. This method can help identify different types of stress, such as physical, cognitive, or emotional stress, and aid in handling them according to psychological conditions. Stress-type classification will be performed using a Multi-Layer Perceptron (MLP) algorithm. The study involved 20 subjects who underwent 3 measurement sessions with different stress stimuli to increase their stress levels. The study results demonstrated that cvxEDA method successfully identified and separated the SC components. The phasic component (SCR) was found to describe the subjects’ physiological responses to the stressors. The MLP model was used to classify 3 different stress types and achieved better accuracy value in the classification of cognitive stress with an accuracy of 82% and an F1-score of 80.13% for cognitive stress. This research provides implications in the form of the potential for early detection of stressful situations so that it can provide treatment according to the stress-type.