Co-infection prevalence of Covid-19 underlying tuberculosis disease using a susceptible infect clustering Bayes Network

Abstract

The adoption of data mining processes is urgently needed due to the everyday generation of large amounts of data at an accelerated rate. The current advancement in the area of data analytics and data science has ushered in a new paradigm shift in the use of machine learning and softcomputing approaches to a new paradigm to render a more beneficial approach in constructing algorithms that can effectively and efficiently assist expert systems to yield new insight to practitioners – to ensure comprehensive decisions on the underlying tuberculosis disease to potential problematic cases. This study explored spatial medical data in disease diagnosis to effectivevly and efficiently handle problematic cases of Tuberclulosisin Nigeria. Bayesian Network algorithm was used to predict potential cases in patients with covid-19 (and other underlying health issues) vis-à-vis its co-prevalence rate with Tuberculosis with data retrieved from the epidemiology laboratoryof the Asaba Federal Medical Centre, Delta State. Training and test versions of the data set were separated. Constructed model yields high prediction compared to previous studies in forecast of the prevalence co-infection rate. Results generated show that the confusion matrix model had sensitivity of 0.81, specificity 0.08, and prediction accuracy of 0.937 for data not originally used to train.