bioRxiv Science⌕ Search

Biology subjects

Patel, D. P.

Publications and source records attributed to Patel, D. P..

2 recordsLinked to original sources
bioRxiv2024

Comparative assessment of machine learning algorithms to predict severity of disease in COVID-19 patients based on eight cofactors

Machine learning is one of the important tools to diagnose and predict the diseased state accurately and effectively. The COVID-19 pandemic caused due to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has become one of the most researched healthcare topics worldwide. Machine learning algorithms can find efficient and reliable ways to predict the COVID-19 from vast amounts of existing health care data, allowing faster, effective, and more accurate diagnosis with lower risk based on the symptoms. Based on the countrywide data published by the Israeli Ministry of Health, we propose a system that detects COVID-19 instances using simple variables. The COVID-19 dataset used in the study consisted of 278848 patients samples with five different symptoms, namely cough, fever, sore throat, shortness of breath, and headache, apart from other basic information like age, gender, and test indication excluding confirmed COVID-19 result. The data was analyzed using traditional supervised machine learning algorithms namely, Decision tree, Support vector machine, Random Forest, Logistic regression, k-nearest neighbor, and Naive Bayes based on eight cofactors with high accuracy rate ([≥] 0.9450). Apart from Support vector machine, all other algorithms displayed better performance based on the AUC score calculated using the receiver operator characteristic (ROC) curve. This study also highlights the significant differences between precision, recall and accuracy for each model. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=97 SRC="FIGDIR/small/617884v1_ufig1.gif" ALT="Figure 1"> View larger version (19K): org.highwire.dtl.DTLVardef@33bb11org.highwire.dtl.DTLVardef@3e93aaorg.highwire.dtl.DTLVardef@508f76org.highwire.dtl.DTLVardef@faa9e2_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOGraphical AbstractC_FLOATNO C_FIG

bioinformatics↗
bioRxiv2023

Prediction of Heart Disease and Survivability using Support Vector Machine and Naive Bayes Algorithm

PurposeIn the present work, we examined the outcomes and accuracy of the Support vector machine (SVM) and the Naive Bayes algorithms on a dataset, to predict whether the patient has heart disease or not, and the patients survival prediction status. MethodThe machine learning procedures were developed using the clinically validated datasets with sixteen attributes from the University of California, Irvines Centre for Machine Learning, and Intelligent Systems. Confusion matrix was used to visualise the accuracy, recall, precision, and error of the models. Statistical analysis was done to prove the model accuracy using the receiver operating characteristic (ROC) curve and area under the curve (AUC). ResultsThe proposed method of heart disease prediction using Naive Bayes had 87 % accuracy. The accuracy for heart survivability models using SVM and Naive Bayes were 88 % and 93 %. The model efficiency for heart survivability using ROC curve with AUC 0.93 for Naive Bayes and AUC 0.91 for SVM. ConclusionSuch prediction systems can help the medical sector to save energy, cost, and time by providing more efficient techniques to forecast decisions with high accuracy. This study will enable the statisticians and researchers to select more efficient and accurate machine learning algorithms to achieve better prediction of the "cardiovascular disease".

bioinformatics↗