Experimental Analysis of Region based Covid-19 Cases Classification using Clustering Strategies

Abstract

Many activities are performed to control the present world pandemic called Corona or else Covid-19 virus, in which all countries are suffered a lot to control such virus with huge efforts. But the resulting to such cases are not properly concluded with a solution to affection and causes a huge death rate around the world. In this paper, a new clustering-based classification strategy is introduced to provide a detailed solution to identify the covid-19 cases, specifically in Saudi Arabia. Because the Covid-19 affection range is too high in Saudi Arabia as well as the situation is too complex to overcome the affection range and wealth of the nation back. This paper provides a good solution with simple and instant processing clustering approach to classify the Saudi Arabia regions, which is affected seriously by Covid-19. This paper introduces a novel clustering approach called "Improved k-Means Clustering Logic (IkMCL)", which is initiated from regular k-means approach but the classification accuracy is improved and the header-based classification strategies are accurate in the proposed approach. In this paper, a group-based classification strategy is followed by the IkMCL and provide the accuracy level as high as the best in outcome. Each header is defined by means of regions, for example states in the country and the groups are categorized into three levels under the header regions such as Low, Medium and High. These categorizations are coming under the ranges of affection over the regions in Saudi Arabia. The pictorization over the following figure, Fig.1 clearly shows the region-based classification view in clear manner. The proposed investigation depicted a straightforward and quick way to deal with screen the Covid-19 pestilence at the territorial and commonplace level and these findings, as of now, offered a preview of the scourge, which could be useful to layout the chain of command of requirements at the sub-national level. Nonetheless, the coordination of the proposed methodology with additional markers and qualities could improve our findings, likewise permitting the application to different settings and with extra points.      The outcome scenario proves the accuracy levels, the effectiveness of proposed system classification, clustering accuracy and grouping purpose in detail with the help of Improved k-Means Clustering Logic.