Predicting Faculty Stress with Machine Learning: Combining Wearable Data and Sentiment Analysis

Abstract

Stress in academic professionals can significantly impact their health and performance. This study proposes a novel approach to stress detection among faculty members using a combination of machine learning, sentiment analysis, and real-time data collection through smart devices. A sample of 500 faculty members from 10 different colleges in Jammu District was selected, and both subjective (survey-based) and objective (physiological data from wearable devices) data were collected. The physiological data from devices like smartwatches and fitness trackers were combined with survey responses to create a comprehensive stress profile. The study uses various machine learning algorithms, including Naive Bayes, Support Vector Machine (SVM), and Random Forest, to predict stress levels based on collected data. The results demonstrate that integrating machine learning with real-time data from smart devices improves the accuracy of stress detection, offering valuable insights for stress management interventions in educational settings.