Graph-Based Clustering Algorithm for Social Community Transmission Control of COVID-19 During Lockdown

Behera, V. N. J.; Ranjan, A.; Reza, M.

This paper proposes a system to model the spread of COVID-19. This system will work in post lockdown conditions, when the only mode of travel, is by road. It defines impact measures, that state the severity of potential disease spread, in a specific area. These impact measures are calculated based on existing hotspots, and are clustered into regions of varying danger-levels, using a graph clustering algorithm. Using this method, it can be predicted where more lenient measures may be taken, and which areas are less prone to the virus spread. There exist other methodologies to model the spread of viruses, but most overlook the spatial nature of viruses. The proposed system focuses on this limitation. Specifically, it focuses on preventing the virus spread, from a geographical point of view. Since the virus spread depends entirely on contact, regions near existing hotspots may potentially become new hotspots. The entire country is first visualized as a weighted graph of regions, at an appropriate administrative level, such as districts. The weights of the nodes are the number of active cases, and the weights of the edges are the geographical distances between those nodes. This graph is connected based on a distance threshold. The impact measure tells the impact of a region, on nearby regions, and the danger value tells the transmission possibility, between separate regions. Using this data, potential hotspots are easily identified. This data will help administrative bodies, to make more fine-tuned lockdown restrictions, based on the impact measures. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.