The Prediction of Mental Stress by Utilizing Supervised and Unsupervised Machine Learning Techniques

Abstract

Mental Stress is the most prominent and pertinent issue nowadays in people all across the globe and can be considered a disease that needs to be addressed on time to avoid suicides. Artificial Intelligence plays a major role in resolving this societal issue by predicting the stress level. In this paper, a hybrid model using various Machine Learning algorithms is constructed to predict the stress based on some crucial parameters.  The constructed hybrid model is hyper-tuned with Adaboost and Bagging techniques and the model's accuracy was raised to 96.4%. The stacked model accurately predicts the stress level and can provide concrete insights into the reasons that are highly responsible for stress among people of all ages.