Towards Mental stress Detection in University Students Based on RBF and Extreme Learning Based Approach

• Published in: 2024 International Conference on Electronics, Computing, Communication and Control Technology (ICECCC) pp. 1 - 6
• Main Authors: Anand, Taruna, Shabbir Alam, Mohammad, Akila, D., Yadav, Manjushri Janardan, Mavliya, Anil Kumar, Kumar, A. Saran
• Format: Conference Proceeding
• Language: English
• Published: IEEE 02.05.2024
• Subjects: Accuracy; Bidirectional Long Short-term Memory (Bi-LSTM); Computational modeling; Feature extraction; Heart; Human factors; Mental Stress Prediction; Noise; Post-Traumatic Stress Disorder (PTSD); Training
• DOI: 10.1109/ICECCC61767.2024.10593904

A lot of individuals associate stressful situations with bad experiences. It is concerning that mental health issues are so common, particularly among younger people. The once carefree generation is now under a great deal of stress. Modern society's elevated stress levels are associated with an array of serious health issues, such as an increased risk of suicide, heart disease, depression, and stroke. Within the scope of our research, we are attempting to quantify the degree to which students experience internet-related anxiety in the seven days leading up to an exam. Model training, feature selection, and preprocessing are all tremendously dependent on sequencing. Images with consistent size and intensity can be achieved by removing noise and blinks during this preprocessing step. Classifier performance improvement via feature selection aims to normalize and increase accuracy by utilizing favorable attributes. Feature selection is crucial during the training of a StressNet-C-BiLSTM model. When compared to two state-of-the-art algorithms, BiLSTM and CNN, the suggested technique outperforms them. The results demonstrated a significant improvement, with an accuracy rate of 97.58%.