Development of Stress Detection and Management System Using Machine Learning Algorithm for Sulu State College: A Proposed Framework for Mitigating Stress Related Diseases
Main Article Content
Abstract
Stress significantly impacts academic performance, mental well-being, and overall student success in higher education institutions. Sulu State College faces challenges in addressing stress among its students and staff due to limited technological interventions. This study aimed to develop and evaluate a comprehensive Stress Detection and Management System using machine learning algorithms, EMG sensors, and DHT11 sensors for the Sulu State College community. The system integrated intramuscular EMG sensors, DHT11 temperature and humidity sensors, Arduino Uno microcontrollers, and machine learning algorithms (Logistic Regression and K-Nearest Neighbors). Data were collected from 25 respondents (18 staff, 7 students). The system was evaluated using a five-point Likert scale by three experts assessing functionality, usability, reliability, and performance. The Perceived Stress Scale (PSS) was administered to measure perceived stress levels. The predictive model achieved 99% accuracy in stress prediction. Expert evaluations yielded general weighted means of 4.77778 (functionality), 4.66667 (usability), 4.5 (reliability), and 5.0 (performance), all categorized as "Very Acceptable." Survey results indicated that 20 out of 25 respondents (80%) experienced moderate stress levels. User satisfaction scored a mean of 4.5 out of 5. The developed Stress Detection and Management System effectively detects and manages stress among students and staff, demonstrating high functionality, usability, reliability, and performance. The system's personalized interventions based on physiological data provide a proactive approach to stress management in academic settings.