Comparison of Machine Learning Algorithms for Predicting Patients with Suspected COVID-19

Abstract

The coronavirus disease (COVID-19) outbreak has infected millions of people, causing a high death rate worldwide. Patients suspected of having COVID-19 are transferred to different health facilities, which has caused a saturation in care, for which it is necessary to have a prediction model to classify patients at high risk of clinical deterioration. The objective of the research was to compare classification algorithms based on automatic learning machines, for the prediction of clinical diagnosis in patients with COVID-19. 1000 records of patients with suspected SARS-CoV-2 infection who were admitted by the emergency service in health establishments in Peru were collected. After pre-processing the data and engineering the attributes, a sample of 700 records was determined. Models were designed and algorithms were compared: Logistic Regression, Support Vector Machine, Nearest Neighbors, Decision Tree, Random Forest, and Navie Bayes. The evaluation of the results of each algorithm was carried out using Accuracy, precision, sensitivity and Chohen's Kappa to know the degree of agreement between the prediction by the learning machine and the results of reality, that is, to what extent both results agree in their measurement. The algorithm that presented the best results was the Support Vector Machine and Random Forest, which predicted the patients with an accuracy of 97%, and Cohen's Kappa of 0.95, with figures higher than the other models evaluated.

Predicting Student Dropout based on Machine Learning and Deep Learning: A Systematic Review

“…This issue entails economic, social, and educational consequences for the stakeholders in the global education system, ranging from the psychological impact on students to the management challenges faced by government entities 4,5 . To address the problem, predicting and managing early signs of student dropout is relevant [6][7][8][9] . This will enable educational institutions to act promptly, implementing preventive and proactive measures to address the issue and reduce the dropout rate 10 .…”