STRESS DETECTION USING MACHINE LEARNING

We all know that stress today is one of the biggest problems in society, and affects our health indirectly, both physically and mentally. Stress is harmful but if recognized timely can also be prevented and properly handled. This paper provides an overview of the new emerging field of Stress Detection using Machine Learning techniques. A new and intriguing recent stream of research with machine learning — which enables the analysis of vast datasets and recognition of non-linear trends — has been the detection of stress. Our method makes use of a individual’s physiological, behavioral and environmental signals and infer their stress levels. For example, Using machine learning algorithms, we can train the models on various features related to stress like age, blood pressure, heart rate of the person to predict whether the person is under stress or not. It may have qualitative characteristics like gender, categories of occupation, or amount of stress. For the classification of human stress level using labeled data, various models can be implemented such as decision tree, random forest, KNN, logistic regression. This abstracts also pointed out the challenges and opportunities in applying ma- chine learning techniques for stress detection. Issues like the need of Huge and Diverse data-sets, moral problems or the chance of model’s bias. Since stress is a complex issue, it must be understood in order to help tackle this problem and with the use of technology, software in the case of its use by individuals and communities to manage the stress and with a combination of these two this is a new and emerging area of research and application to machine learning in the world and using technology by persons and individuals makes the stress detection get a good result for a good life until the main task is done.