Scholarlink

<p>Stress is a pensive problem across the globe, and this issue has enormous impacts on mental and physical health. Owing to long-term exposure to stress, serious health problems may arise. Therefore, its timely detection is supportive for controlling stressful periods. Accordingly, various multimodal (MM) signal-based approaches are expansively explored, as stress seriously impacts the structure and functioning of the brain. These techniques developed for detecting stress require enhancement regarding reliability and prediction accuracy. Therefore, this research presents a new model termed multigate long short-term memory (LSTM) for stress detection utilizing MM signal. Initially, MM signals are obtained to process stress detection. The considered MM signals are acquired from four sensors namely a temperature sensor, electrodermal activity (EDA) sensor, photoplethysmograph (PPG) sensor, and average heart rate (HR) as well as a 3-axis accelerometer sensor. After that, features from four sensors are individually extracted to obtain output- 1, output- 2, output- 3, and output- 4. Lastly, stress detection is performed utilizing Multigate-LSTM by taking four feature-extracted outputs. In addition, Multigate-LSTM obtained minimum values of mean absolute percentage error (MAPE) of about 0.172, mean square error (MSE) of about 0.082, and root mean square error (RMSE) of about 0.286.&#xa0;</p>

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

Multigate Long Short-Term Memory-Based Stress Detection from Multimodal Signal

  • 期刊论文

  • 发表日期：

    2025年5月10日

查看全文
  • Meenalakshmi M.,
  • Valarmathi K.

摘要

Stress is a pensive problem across the globe, and this issue has enormous impacts on mental and physical health. Owing to long-term exposure to stress, serious health problems may arise. Therefore, its timely detection is supportive for controlling stressful periods. Accordingly, various multimodal (MM) signal-based approaches are expansively explored, as stress seriously impacts the structure and functioning of the brain. These techniques developed for detecting stress require enhancement regarding reliability and prediction accuracy. Therefore, this research presents a new model termed multigate long short-term memory (LSTM) for stress detection utilizing MM signal. Initially, MM signals are obtained to process stress detection. The considered MM signals are acquired from four sensors namely a temperature sensor, electrodermal activity (EDA) sensor, photoplethysmograph (PPG) sensor, and average heart rate (HR) as well as a 3-axis accelerometer sensor. After that, features from four sensors are individually extracted to obtain output- 1, output- 2, output- 3, and output- 4. Lastly, stress detection is performed utilizing Multigate-LSTM by taking four feature-extracted outputs. In addition, Multigate-LSTM obtained minimum values of mean absolute percentage error (MAPE) of about 0.172, mean square error (MSE) of about 0.082, and root mean square error (RMSE) of about 0.286.