Related Experiment Video
Updated: Sep 5, 2025
08:20
Author Spotlight: AI-Driven Trypanosome Species Detection from Microscopic Images
Published on: October 27, 2023
1.6K
Detection of COVID-19 Patients Using Machine Learning Techniques: A Nationwide Chilean Study
Pablo Ormeño1, Gastón Márquez2, Camilo Guerrero-Nancuante3
1Escuela de Ingenieria y Negocios, Universidad de Viña del Mar, Viña del Mar 2520000, Chile.
Summary
Machine learning models, including support-vector machines, can effectively classify COVID-19 patients using symptom data from Chile's Epivigila surveillance system. Support-vector machines demonstrated superior performance in this analysis.
Area of Science:
- Epidemiology
- Machine Learning
- Public Health
Background:
- Chile's Epivigila system contains over 17 million patient records, offering valuable data for COVID-19 analysis.
- Classifying symptoms and comorbidities in large datasets is challenging for healthcare professionals.
Purpose of the Study:
- To compare machine learning techniques for COVID-19 patient classification.
- To identify which symptoms and comorbidities are associated with COVID-19 infection.
Main Methods:
- Comparison of Support-Vector Machine, Decision Tree, and Random Forest algorithms.
- Evaluation using precision, recall, accuracy, F1-score, and AUC metrics.
- Utilized a 10% sample of confirmed COVID-19 patients from the Epivigila database.
Main Results:
- Support-vector machine outperformed Decision Tree and Random Forest in recall, accuracy, F1-score, and AUC.
- Machine learning efficiently processes and classifies large volumes of health data.
- Faster healthcare decision-making is facilitated by these techniques.
Conclusions:
- Support-vector machine is a highly effective tool for COVID-19 classification using epidemiological data.
- Machine learning enhances the analysis of large-scale health surveillance data.
- Automated classification accelerates critical healthcare decisions during pandemics.
