A Reconstructive Model for Identifying the Global Spread in a Pandemic

Abstract

With existing tracing mechanisms, we can quickly identify potentially infected people with a virus by choosing everyone who has come in contact with an infected person. In the presence of abundant resources, that is the most sure-fire way to contain the viral spread. In the case of a new virus, the methods for testing and resources may not be readily available in ample quantity. We propose a method to determine the highly susceptible persons such that under limited testing capacity, we can identify the spread of the virus in a community. We determine highly suspected persons (represented as nodes in a graph) by choosing paths between the infected nodes in an underlying contact graph (acquired from location data). We vary parameters such as the infection multiplier, false positive ratio, and false negative ratio. We show the relationship between the parameters with the test positivity ratio (the number of infected nodes to the number of suspected nodes). We observe that our algorithm is robust enough to handle different infection multipliers and false results while producing suspected nodes. We show that the suspected nodes identified by the algorithm result in a high test positivity ratio compared to the real world. Based on the availability of the test kits, we can run our algorithm several times to get more suspected nodes. We also show that our algorithm takes a finite number of iterations to determine all the suspected nodes.

Dibakar Saha would like to acknowledge the Science and Engineering Research Board (SERB), Government of India, for financial support under the NPDF scheme (File Number: PDF/2018/000633).