Newcastle disease virus clustering based on swarm rapid centroid estimation

Abstract

Newcastle disease is considered to be one of the most important serious infectious poultry disease. The work introduced in this paper addresses the problem of clustering Newcastle disease dataset obtained from the National Center for Biotechnology Information GenBank (NCBI). A lightweight swarm clustering algorithmcalled Rapid Centroid Estimation (RCE) is applied in the clustering task. However, the best number of clusters is selected using silhouette measure. Hence, RCE is compared with K-means for the same number of clusters. The experiment shows that the external quality measures (purity, entropy, rand index, precision, recall and F-measure) of the RCE clustering technique outperform the ones of the K-means.