Machine Learning Method to Discover the Novel Association of Vaccine and Vitamin A Supplement on Covid 19 Infection: A Biclustering Approach

ABSTRACT

This article attempts to discover the novel association of vaccines (administered for other diseases) and vitamin A supplement coverage across different countries and the spread of COVID-19.This relation may be affected by several unknown factors as the disease spreads over regions with diverse economic, cultural and climatic conditions. In this situation, relevant associations between features may exist for a subset of countries rather than the entire set of countries.We use a machine learning method named RelDenClu to identify the effects of vaccines or vitamin supplements on the spread of COVID-19. RelDenClu is a non-linear relation based biclustering technique. From the experiment, it was found that countries providing the Tetanus-Toxoid vaccine have lower COVID-19 infection rates. The performance of the proposed technique is also compared with decision tree, LASSO and CBSC and it is noticed that all these methods discover the same association (i.e., countries with a higher rate of administering Tetanus-Toxoid vaccine show lower COVID-19 infection). However, the proposed method can discover the said association in less computation time. Additionally, the proposed method is general enough to be applied to other datasets for finding associations between different local factors affecting the proliferation of any disease across different regions. © 2021 IEEE.