Multivariate Analysis and Machine Learning - Mortality Predictions In COVID-19 Patients from Comorbidity, Demographic and Laboratory Findings

Q: Which machine learning algorithms were compared for COVID-19 mortality prediction?

The study compared six supervised algorithms: KNN, Naive Bayes, SVM, decision tree, random forest, and logistic regression, evaluated with 10-fold cross-validation.

Q: What variables were identified as top mortality predictors?

The top predictors reported were Mean Artery Pressure (MAP), stroke, age, and IL6, with statistically significant associations.

Q: How well did the best model perform and what evaluation metrics were used?

Logistic regression was reported as the best model, achieving about 81% ROC. Performance was assessed using precision, recall, F-measure accuracy, and ROC AUC.

Aug 5, 2026

Upload by Valentina

Study objective focuses on predicting mortality risk among COVID-19 patients using machine learning models supported by multivariate analysis of clinical, demographic, and laboratory variables. A retrospective dataset of 4711 cases was assembled from public data and evaluated with six supervised algorithms (KNN, Naive Bayes, SVM, decision tree, random forest, logistic regression) under 10-fold cross-validation. Model quality was measured via precision, recall, F-measure accuracy, and ROC AUC. Logistic regression achieved the strongest ROC performance and identified MAP, stroke, age, and IL6 as key predictors.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages12

Reading Time30 min

AccessFree to download

Abstract

Introduction

Data mining for healthcare decision support

Methodology

Dataset collection and description

Data mining techniques and algorithms

Logistic Regression (LR)

Multivariate Analysis and Machine Learning: Mortality Predictions In COVID-19 Patients from Comorbidity, Demographic and Laboratory Findings Husnul Khuluq1, Prasandhya Astagiri Yusuf 2, Dyah Aryani Perwitasari 3, Abdul Fadhil 4 1 Dept. of Pharmacy, Faculty of Health Sciences, Universitas Muhammadiyah Gombong 2 Dept. of Medical Physiology of Biophysis / Medical Technology Cluster IMERI, Faculty of Medicine, Universitas Indonesia 3 Faculty of Pharmacy, Universitas Ahmad Dahlan 4 Dept. of Electrical Engineering, Technical Faculty, Universitas Ahmad Dahlan Received: 24-June-2023 Revised: 27-July-2023 Accepted: 21-August-2023 Abstract Objective: COVID-19 Patients were constantly at a risk of death. It has been demonstrated that the utilization of machine learning (ML) algorithms could be a possible strategy for prediction mortality. Aim: This study aimed to analysis six Machine Learning (ML) algorithms in an multivariate analysis to identify key clinical, demographic and laboratory finding to predict mortality in COVID-19 pandemic Materials and methods: This retrospective study consisted of persons-under-investigation for COVID-19. Dataset taken from data science community ([kaggle.com](kaggle.com)), predictive models of mortality were constructed and compared using six supervised machine learning algorithms: KNN, naivebayes, SVM, decision tree, random forest and logistics regression using 10-fold cross-validation and multivariate analysis. The performance of algorithms was assessed using precision, recall, Fmeasure accuracy and area under the receiver operating characteristic curve (ROC) . The Waikato Environment for Knowledge Analysis (WEKA) version 3.8.6 for analysis. Multivariate analysis using Logistic regression were used to predict mortality. Results: A total of 4711 patients were included in the analysis. The top 4 mortality predictors were Mean Artery Pressure (MAP) (p<0.001; OR 17.071(12.233-23.820), stroke (p<0.001;OR 3.499(1.883-6.503), Age (p<0.001;OR 3.23(2.716-3.830), IL6 (p<0.001; OR 2.03(1.512-2.725. Logistic regression was the best ML algorithms predicted mortality with 81% ROC. Conclusion: This study identifies important independent clinical variables that predict COVID-19 infection-related mortality. The prediction method is helpful, easily improved, and easily retrained with new data. This method can be applied right away and may help front-line doctors make clinical decisions in situations where there are limited resources and time. Keywords: big data study, data mining research, machine learning algorithm, prediction models 1. INTRODUCTION Clinical Severe Acute Respiratory Syndrome Coronavirus (SARS-CoV-2), the responsible agent of novel coronavirus (COVID-19 or 2019-nCoV), appeared in late 2019 and likely to come from Hubei Province, China called Wuhan[1][2] It is suspected that COVID-19, which is quickly spreading in humans, was initially originated from bats and likely spread to humans through intermediate hosts, the raccoon dog (Nyctereutes procyonoides) and palm civet (Paguma larvata) [3][4] The earliest symptom of SARS-COV-2 were fever, coughing, and shortness of breath, which frequently matched the flu. [2] Since then, COVID-19 advanced to a critical stage and spread globally, infecting numerous people. Human-to-human transmission of COVID-19 from infected patients with moderate symptoms has also been documented. [5]. Nevertheless, no drug or vaccine has been clinically shown to cure COVID-19 pandemic, so other non-clinical or non-medical therapeutic techniques, such as data mining techniques, machine learning, and expert systems, among other artificial intelligence techniques, are needed to contain and prevent further outbreak. Data Mining (DM) is a sophisticated AI methods for finding new, practical, and reliable hidden patterns or knowledge from datasets. [6] The method identifies connections, information, or patterns between the datasets in multiple or a specific datas

FAQs

Which machine learning algorithms were compared for COVID-19 mortality prediction?

The study compared six supervised algorithms: KNN, Naive Bayes, SVM, decision tree, random forest, and logistic regression, evaluated with 10-fold cross-validation.

What variables were identified as top mortality predictors?

The top predictors reported were Mean Artery Pressure (MAP), stroke, age, and IL6, with statistically significant associations.

How well did the best model perform and what evaluation metrics were used?

Logistic regression was reported as the best model, achieving about 81% ROC. Performance was assessed using precision, recall, F-measure accuracy, and ROC AUC.

Aug 5, 2026

Upload by Valentina