Clustering Study of Indian States and Union Territories Affected by Coronavirus (COVID-19) using k-means Algorithm Poompaavai A., Manimannan G.* Assistant Professor, Department of Statistics, Appllo College of Arts and Science, Chennai, India *Corresponding author email id: manimannang@gmail.com
Online published on 21 May, 2020. Abstract In this research paper an attempt is made to identify the affected Indian states and Union Territories by Corana Virus Disease-19 (COVID-19) using clustering method based on the secondary sources of data that was collected from Indian Health and Family welfare Organization till March 24, 2020. The worldwide spread of Severe Acute Respiratory Syndrom (SARS) epidemic has clearly shown the significance of considering long-range moving networks in accepting emerging diseases epidemics. It is spread through human to human. Many research scholars reviewed and analysed about this virus. Most of the scholars used Descriptive level statistics for their report in early studies. Very few scholars used higher end machine learning statistical methos for their research from 1950 onwards. In this research paper, the researchers tried to cluster the data based on silhouette distance measure using k-means clustering algorithm. It produced effective result and visualises their result in a simple manner. Conclusion: Three groups of clusters has formed. Maharashtra and Kerala in cluster C2 based on Silhouette distance measure and this cluster is labelled as High COVID 19. Cluster C1 consists of Andhra Pradesh, Bihar, Chhattisgarh, Haryana, Himachal Pradesh, Madhya Pradesh, Odisha, Puduchery, Tamil Nadu, Chandigarh, Jammu and Kashmir, Ladakh, Uttarakhand, and West Bengal states are labelled as Moderate COVID-19. Cluster C3 consists of the following, they are Delhi, Gujarat, Karnataka, Punjab, Rajasthan, Telangana and Uttar Pradesh. In addition, the visualization of all the parameters is depicted in this research paper. Top Keywords COVID-19, Indian States, k-means clustering, Silhouette distance and visualization. Top |