AI-driven health analysis for emerging respiratory diseases: A case study of Yemen patients using COVID-19 data.

Journal: Mathematical biosciences and engineering : MBE
PMID:

Abstract

In low-income and resource-limited countries, distinguishing COVID-19 from other respiratory diseases is challenging due to similar symptoms and the prevalence of comorbidities. In Yemen, acute comorbidities further complicate the differentiation between COVID-19 and other infectious diseases. We explored the use of AI-powered predictive models and classifiers to enhance healthcare preparedness by forecasting respiratory disease trends using COVID-19 data. We developed mathematical models based on autoregressive (AR), moving average (MA), ARMA, and machine and deep learning algorithms to predict daily confirmed deaths. Statistical models were trained on 80% of the data and tested on the remaining 20%, with predicted results compared to actual values. The ARMA model demonstrated promising performance. Additionally, eight machine learning (ML) classifiers and deep learning (DL) models were utilized to identify COVID-19 severity indicators. Among the ML classifiers, the Decision Tree (DT) achieved the highest accuracy at 74.70%, followed closely by Random Forest (RF) at 74.66%. DL models showed comparable accuracy scores, around 70%. In terms of AUC-ROC, the kernel Support Vector Machine (SVM) outperformed others, achieving 71% accuracy, with precision, recall, F-measure, and area under the curve values of 0.7, 0.75, 0.59, and 0.72, respectively. These findings underscore the potential of AI-driven health analysis to optimize resource allocation and enhance forecasting for respiratory diseases.

Authors

  • Saleh I Alzahrani
    Biomedical Engineering Department, College of Engineering, Imam Abdulrahman Bin Faisal University, PO box 1982, Dammam 31451, Saudi Arabia.
  • Wael M S Yafooz
    Computer Science Department, Taibah University, Saudi Arabia.
  • Ibrahim A Aljamaan
    Biomedical Engineering Department, College of Engineering, Imam Abdulrahman Bin Faisal University, PO box 1982, Dammam 31451, Saudi Arabia.
  • Ali Alwaleedi
    Department of Epidemiology and Public Health, College of Medicine, Aden University, Aden, Yemen.
  • Mohammed Al-Hariri
    Department of Physiology, College of Medicine, Imam Abdulrahman Bin Faisal University, PO box 1982, Dammam 31451, Saudi Arabia.
  • Gameel Saleh
    Biomedical Engineering Department, College of Engineering, Imam Abdulrahman Bin Faisal University, PO box 1982, Dammam 31451, Saudi Arabia.