Tunicate Swarm-Based Black Hole Entropic Fuzzy Clustering for Data Clustering using COVID Data

Abstract

The corona-virus (COVID-19) pandemic outburst from China has infected number of peoples and causes many deaths. In addition, the count of infections and deaths rates augment quickly. The adaption of data mining for performing recognition of infectious patterns is utilized for analyzing he spread patterns of COVID-19 infection. The grouping of clinical information is major method for determining the concealed instances from large clinical dataset. The clustering assist to group the information obtained from different gatherings. This paper devises a novel methodology, namely Tunicate swarm algorithm-based Black-hole entropic fuzzy clustering (TSAbased BHEFC) for clustering the COVID data. The clustering is performed using a Black Hole Entropic Fuzzy Clustering (BHEFC) technique. The weighted coefficients equivalent to cluster centers is optimally found using Tunicate swarm algorithm (TSA). Here, the log transformation is adapted for transforming the data in order to make it suitable for further processing. In addition, significant features are selected using Pearson Correlation coefficient. Subsequently, the chosen features are fed to clustering phase, where the clustering of COVID patients are performed with proposed TSA-based BHEFC. The proposed TSA-based BHEFC algorithm outperformed with maximal accuracy of 95.061%, maximal jaccard coefficient of 90.852% and maximal dice coefficient of 90.420%.