9. Fog-IoT Environment in Smart Healthcare: A Case Study for Student Stress Monitoring

Fog computing disseminates computing system which incorporates the cloud computing model to fully support the vision of internet of things (IoT). In the course of the most recent couple of years, the internet of things (IoT) opens the portal to developments that encourage communication among things as well as among people known as the man to machine (M2M) interface. Concentrating on medicinal services space, IoT devices, for example, therapeutic sensors, visual sensors, cameras, as well as remote sensor systems, are driving the developmental pattern. Toward this way, the part anticipates strengthening the amalgamation of fog computing in the medicinal services area. Convinced by the equivalent creative methods, our work features the latest IoT-aware student-centered stress management system for student stress indexing in a specific context. The work proposes to utilize the temporal dynamic Bayesian network (TDBN) model to depict the event of stress as conventional or sporadic by readings through physiological means congregated from medicinal devices at the fog layer. Constructed from four parameters, especially leaf node confirmations, outstanding tasks at hand, context, and understudy well-being quality are employed for the stress computation, and decisions are made well into the shape of a warning generator equipment with provision of moment-sensitive information to caregivers or respondents. Experimentation is aimed on both fog and cloud layers on stress-related datasets that illustrate the usefulness and accuracy of the TDBN model in our proposed system. The final experiments bear witness that the BBN classifier overweighed the group by attaining an accuracy value of 95.5% and specificity of 97.3%, whereas J48, Random forest, and SVM have accomplished an exactness of 85.2%, 87.9%, and 90.8%, separately. However, if sensitivity and f-measure would occur, the BBN classifier beats other classifier models individually with 95.5% and 92.9% values for the same. Also, we evaluated our proposed method with seven states of the artworks, and again, our method leads the list in terms of its promised performance. The work also offers a gentle touch in the literature review form on the recent novel techniques and methods, including deep learning for complex heterogeneous healthcare sensor data, which act as a supporting hand for fog computing.