A Deep Learning Approach for Diagnosis of COVID-19 Infection and Its Related Factors: A Population-Based Study

    September 2023 in “ JP Journal of Biostatistics ”
    Abolfazl Payandeh, Habibollah Esmaily, Masoud ‎Salehi, Seyed Mahdi Amir Jahanshahi, Maryam Salari, Seyed Ali Alamdaran, Ahmad Bolouri
    TLDR The random forest model effectively helps diagnose COVID-19 using key factors like age and symptoms.
    This study utilized deep learning (DL) algorithms to diagnose COVID-19 and identify significant related factors in a population-based study involving 10,862 individuals from 35 hospitals in Khorasan-Razavi, Iran. Among the DL models tested, the random forest (RF) algorithm demonstrated superior performance with a sensitivity of 66%, specificity of 95%, precision of 88%, accuracy of 85%, and an AUC of 74%. Key predictors for COVID-19 detection included age, SpO2, reception season, CT results, contact history, sex, and fever. The RF model shows promise for aiding clinical decision-making and resource allocation, though it requires external validation in larger, more diverse populations.
    Discuss this study in the Community →

    Research cited in this study

    1 / 1 results