Determinación automática del color del semáforo Mexicano del COVID-19 a partir de las noticias

Miguel Ángel Alvarez-Carmona; Ramón Aranda

doi:10.1590/SciELOPreprints.3834

##article.authors##

Miguel Ángel Alvarez-Carmona CICESE-UT3 https://orcid.org/0000-0003-4421-5575
Ramón Aranda CICEsE-UT3 https://orcid.org/0000-0001-8269-3944

DOI:

https://doi.org/10.1590/SciELOPreprints.3834

Keywords:

COVID-19, natural language processing, textual classification, epidemiological semaphore

Abstract

This paper presents the analysis of textual classification models to automatically determine the Mexican regional epidemiological traffic light through COVID news. A database was collected with 4,270 news items referring to COVID, from June 1, 2020, to March 28, 2021. The label of each news item is the color of the epidemiological traffic light that the Mexican government cataloged in the week of publication of the news Classifiers such as SVM, KNN, Random Forest, and Deep Learning were applied. The results show that it is possible to take advantage of the information published in the news to determine the color of the traffic light up to 4 weeks in advance, obtaining results of up to 0.74 F-measure, which is a competitive result taking into account the imbalance of classes of this task.