Avaliação do Uso de IA para Auxiliar no Diagnóstico de Casos de COVID-19 para Diferentes Surtos

• Felipe C. Ulrichsen UERJ
• Alexandre C. Sena UERJ
• Luís C. M. S. Pôrto UERJ
• Karla Figueiredo UERJ

DOI: https://doi.org/10.5753/sbcas.2022.222467

Abstract

Experience with COVID-19 has indicated that, in high demand situations or at underserved regions, there may be a lack of RT-PCR tests, considered standard for COVID-19 diagnosis. In this context, this work researches algorithms, based on Machine Learning (ML), for diagnosing COVID-19 only from signs and symptoms. In addition to diagnosis, this work also included studies considering outbreak waves and testing (RT-PCR and Rapid Antigen Test). The results indicate that the selected model, based on MLP Neural Networks, has a much more accurate result (70% in 1st the wave) than the one indicated by the Ministry of Health for the Rapid Test (around 25%), and indicates the changes in signs and symptoms during the evolution of the pandemic.

References

Brant, R. (1996). Digesting logistic regression results. The American Statistician, 50(2):117-119.

Breiman, L. (1996). Bagging predictors. Machine Learning, 24:123-140.

Breiman, L. (2001). Random forests. Machine Learning, 45(1):5-32.

de Souza, U., dos Santos, R., Campos, F., Lourenço, K., da Fonseca, F., and Spilki, F. (2021). High rate of mutational events in sars-cov-2 genomes across brazilian geographical regions. Viruses, 13.

Fang, Y., Zhang, H., J., X., Lin, M., Ying, L., Pang, P., and Ji, W. (2020). Sensitivity of chest ct for covid-19: Comparison to rt-pcr. Radiology, 296:115-117.

Hoffmann, M., Kleine-Weber, H., Schroeder, S., Kruger, N., Herrler, T., and Erichsen, S. (2020). Sars-cov-2 cell entry depends on ace2 and tmprss2 and is blocked by a clinically proven protease inhibitor. Cell, 16:271-280.

Kononenko, I. (1997). Overcoming the myopia of inductive learning algorithms with relieff. Applied Intelligence, 7(1):39-55.

Lundberg, S. M. and Lee, S.-I. (2017). A unified approach to interpreting model predictions. In Proc. of the 31st Inter. Conference on Neural Information Processing Systems, NIPS'17, page 4768-4777, Red Hook, NY, USA. Curran Associates Inc.

Mark A. Hall, L. A. S. (1999). Feature selection for machine learning: Comparing a correlation based filter approach to the wrapper. Proceedings of the Twelfth International FLAIRS Conference, 30:4765-4774.

Menni, C. et al. (2020). Real-time tracking of self-reported symptoms to predict potential covid-19. Nature Medicine, 26(7):1037-1040.

Murdoch, T.B., D. A. (2013). The inevitable application of big data to health care. Jama, 309(13)::1351-1352.

Rumelhart, D. J. M. (1986). Parallel distributed processing: explorations in the microestructure of cognition. Cambridge: MIT Press.

Souza, M., Figueiredo, K., Porto, L., and Medronho, R. (2021). Experiências e impacto da pandemia pela Covid-19 no complexo de Saúde UERJ, volume 1. Rio de Janeiro.

Topol, E. (2019). Deep medicine: How artificial intelligence can make healthcare human again. Basic Books, 1.

Wiens, J., S. E. (2018). Machine learning for healthcare: On the verge of a major shift in healthcare epidemiology. Clinical Infectious Diseases, 66:149-153.

Yan, L. (2020). A machine learning-based model for survival prediction in patients with severe covid-19 infection. medRxiv.

Zhang, X. et al. (2020). Biological, clinical and epidemiological features of covid-19, sars and mers and autodock simulation of ACE2. Infect Dis Poverty, 20):99.

Zhu, M. et al. (2021). Molecular phylogenesis and spatiotemporal spread of sars-cov-2 in southeast asia. Public Health, 9.