Diagnosis of COVID-19 through blood sample using ensemble genetic algorithms and machine learning classifier

Purpose This purpose of this study is to perfrom the analysis of COVID-19 with the help of blood samples. The blood samples used in the study consist of more than 100 features. So to process high dimensional data, feature reduction has been performed by using the genetic algorithm. Design/methodolog...

Full description

Saved in:
Bibliographic Details
Published inWorld journal of engineering Vol. 19; no. 2; pp. 175 - 182
Main Authors Doewes, Rumi Iqbal, Nair, Rajit, Sharma, Tripti
Format Journal Article
LanguageEnglish
Published Brentwood Emerald Publishing Limited 15.03.2022
Emerald Group Publishing Limited
Subjects
Accuracy
Ant colony optimization
Blood
Blood tests
Coronaviruses
COVID-19
Dehydrogenases
Disease
Genetic algorithms
Hospitals
Low income groups
Machine learning
Medical imaging
Medical research
Pandemics
Parameter sensitivity
Performance evaluation
Programming languages
Radiation
Recovery (Medical)
Reduction
Sensitivity
Severe acute respiratory syndrome coronavirus 2
Ant colony optimization
Feature selection
SARS-CoV-2
Random forest
Relief
Online AccessGet full text

Cover

More Information
Summary:Purpose This purpose of this study is to perfrom the analysis of COVID-19 with the help of blood samples. The blood samples used in the study consist of more than 100 features. So to process high dimensional data, feature reduction has been performed by using the genetic algorithm. Design/methodology/approach In this study, the authors will implement the genetic algorithm for the prediction of COVID-19 from the blood test sample. The sample contains records of around 5,644 patients with 111 attributes. The genetic algorithm such as relief with ant colony optimization algorithm will be used for dimensionality reduction approach. Findings The implementation of this study is done through python programming language and the performance evaluation of the model is done through various parameters such as accuracy, sensitivity, specificity and area under curve (AUC). Originality/value The implemented model has achieved an accuracy of 98.7%, sensitivity of 96.76%, specificity of 98.80% and AUC of 92%. The results have shown that the implemented algorithm has performed better than other states of the art algorithms.
ISSN:1708-5284
1708-5284
DOI:10.1108/WJE-03-2021-0174