Employing Machine Learning Algorithms to Detect Stress with a Specific Emphasis on Commuting Methods

The regular commute for many individuals could significantly impact their general well-being. The daily commute to work can be linked to chronic stress, which is known to have negative implications on mental health, as well as increased blood pressure, heightened heart rate, and high fatigue. The primary objective of this study is to examine the physiological effects of commuting using machine learning techniques, with a specific emphasis on analysing the impact of different transportation methods. Healthy individuals were recruited to collect various biological signals, such as blood pressure (BP), heart rate, and electroencephalogram (EEG) data. By leveraging multiple machine learning techniques, we examined the effects of different commuting modes, whether short or long. Our findings revealed an increase in objective bio signals following the commute. Furthermore, when comparing stress levels between different commute modes, we observed that driving is more stressful than other modes, like public transport. We obtained highly encouraging outcomes by implementing the support vector machine (SVM) algorithm, which exhibited an impressive accuracy of 93.2%. In comparison, the K-nearest neighbour (KNN) and Naïve Bayes algorithms yielded good accuracy of 87.9%. Similarly, by utilising the PANAS questionnaire, we observed that the positive affect levels were greater before the commute. This suggests that participants demonstrated a higher degree of positivity and enthusiasm towards their work prior to boarding on their commute.