Clustering West Nile Virus Spatio-temporal data using ST-DBSCAN

Abstract

Spatio-temporal data mining has been the talk of the day due to high availability of spatio-temporal data from varied sources in diverse fields. Through many tracking devices, huge amounts of spatio-temporal data are being generated. In epidemiology, diseases, patterns and trends attached can be explored taking advantage of methods such as spatio-temporal clustering to discover new knowledge. In this paper Spatio-Temporal Density Based Spatial Clustering of Applications with Noise (ST-DBSCAN) is implemented and analysed on a public health dataset. Upon the implementation, results are analysed, loopholes spotted and a fuzzy version of ST-DBSCAN is proposed. The method is successfully applied to find spatio-temporal clusters in Chicago West Nile Virus (WNV) surveillance data for the period 2007 to 2017.The drawbacks in the original ST-DBSCAN are identified and solutions are proposed. ST-DBSCAN is an extension of the original Density Based Spatial Clustering of Applications with Noise (DBSCAN).