Smart Diagnostics of COVID-19 With Data-Driven Approaches

Abstract

The traditional assays and diagnostic methods are time-consuming and expensive. As the COVID-19 pandemic is expected to remain for a while, it is demanded to develop an efficient diagnosis system. This chapter is designed to investigate how to incorporate data-driven approaches to the construction of a smart health framework for COVID-19. Topics cover a broad range of smart diagnosis innovations for supporting current assays and diagnostics, such as data analysis for nucleic acid tests, machine learning-based serological signatures identification, medical image classification using deep learning, and decision support system for automatic diagnosis with clinical information. Each topic has been illustrated and discussed throughout methodologies, data collections, experimental designs and results, limitations, and potential improvements. All applicational potentials have been examined with real-world datasets. The findings conclude that big data and AI work for providing insightful suggestions on multiple diagnostic assays and COVID-19 detection approaches.