To search, Click below search items.

 

All Published Papers Search Service

Title

Data Clustering Mapping of Covid-19 Pandemic Based On Geo-Location and Machine Learning

Author

Mustafa Abdul Slaam, Karam Gouda, Ahmad Naguib

Citation

Vol. 22  No. 4  pp. 473-480

Abstract

The spread of the covid-19 virus pandemic is very fast, where was start of the virus spreading from Wuhan City, Hubei Province, China and suddenly spread out widely to almost around the world. Because of the quick spread of the Covid-19 pandemic, the disease's contagious nature, and the delay of vaccine production, the only option is to prevent people from mingling in a mob. Using data mining techniques to clustering hotspots zones for impacted Corona positive patients and narrowing the focus to only those zones can be one of the suitable solution against the spread of pandemic. Up-to-date and reliable information about hotspot zones can help the government efficiently implement the measures by focusing resources on the zones, as well as notify other residents about such hotspot zones. The most typical use of hotspot detection in public health is to identify the outbreaks of diseases. There are many Clustering algorithms used for clustering COVID-19 Pandemic , this research aims to compare between two method of clustering technique DBSCAN (Density-Based Spatial Clustering of Applications with Noise) and K-Means, and evaluate the performances of these clustering algorithms using Silhouette score values for both and elbow method for K-mean , and discuss which one is fit for purpose of clustering the hotspot , to produce a powerful hotspot map that can help the decision makers. The proposed approaches are evaluated using Data Science for COVID-19 (DS4C) dataset. The dataset was retrieved from the official repository of the Korea Centre for Disease Control and Prevention (KCDC). From data testing for 244 patient’s geolocation (Long., Lat.), show that DBSCAN method separate the data to 4 main cluster noise points with eps=0.45 and minimum pts=20, and for K-Means method with k = 4, including all points, as no noise points in K-mean cluster method.

Keywords

DBSCAN, K-Means, Geo-location, COVID-19

URL

http://paper.ijcsns.org/07_book/202204/20220456.pdf