A Supervised Learning-Based Framework for Predicting COVID-19 in Patients

Abstract

The integration of ML and loT can provide insightful details for critical decision making, automated responses, etc. Predicting future trends and detecting anomalies are some of the areas where loT and ML are being used at a rapid rate. Machine learning can help decode the hidden patterns in IoT data. It may complement or replace manual processes in critical areas with automated systems that use statistically derived behavior. In healthcare, wearable sensors used for tracking patient activity have been continuously producing a staggering amount of data. This paper proposes an IoT-based scalable architecture for detecting COVID-19-positive patients and storing and processing such massive amount of data on the cloud. The proposed architecture also employs machine learning algorithms for correct classification of patients. The proposed architecture employs gradient boosting classifier method for early detection of COVID-19 in the patient's body. In order to make the architecture scalable and faster in terms of computational power, the architecture employs cloud computing for data storage.