Pronóstico del COVID-19 en Colombia utilizando Redes Neuronales Recurrentes con celdas de gran memoria decorto plazo y unidades recurrentes cerradas

• Yeison Armando Buitrago López ^[1] ; Luis Alejandro Másmela Caita ^[1]
  1. [1] Universidad Distrital
• Localización: Comunicaciones en Estadística, ISSN 2027-3355, ISSN-e 2339-3076, Vol. 15, Nº. 2, 2022, págs. 16-38
• Idioma: español
• Títulos paralelos:
  • Forecasting of COVID-19 in Colombia using recurrent neural networks with longshort term memory and gated recurrent units
• Enlaces
  • Texto completo (pdf)
• Resumen
  • español

    El 6 de marzo del 2020, el primer caso de COVID-19 fue reportado en Colombia, este virus, declaradocomo una emergencia de salud p ́ublica de importancia internacional ha afectado diferentes sectores.Existe un auge en cuanto al n ́umero de estudios que buscan hacer pron ́osticos en diversos aspectosque tienen que ver con este virus. El presente trabajo muestra los aspectos te ́oricos de las redesneuronales recurrentes y se utilizan para crear una predicci ́on de 60 d ́ıas sobre los casos acumulados,fallecidos acumulados y recuperados acumulados disponibles desde el 6 de marzo del 2020 hasta el6 de marzo del 2022. Redes neuronales con celdas GRU y LSTM junto con las cl ́asicas RNN fueronutilizadas para hacer estos pron ́osticos.

  • English

    On march 6 of 2020, the first case of COVID-19 was reported in Colombia. This virus, declared apublic health emergency of international importance, has affected different sectors. There is a boomin the number of studies that make forecasts in various aspects that have to do with this virus. Thepresent work shows the theoretical aspects of recurrent neuronal networks and his use to create a60-day forecast on cumulative cases, cumulative deaths and cumulative recovered, available frommarch 6 2020 to march 6 2022. Neural networks with GRU and LSTM cells along with the classicRNN were used to make these forecasts.

• Referencias bibliográficas
  • J. Brownlee. Deep Learning for Time Series Forecasting. Machine Learning Mastery, fourth edition, 2018.
  • Y. Buitrago. Pronóstico del COVID-19 en Colombia. https://github.com/YeisonABL/Pronostico_Colombia_COVID19, 2022. [En línea; acceso 6/marzo/2022].
  • U. o. T. Computer Science. Lecture 15: Exploding and Vanishing Gradients. https://www.cs.toronto.edu/~rgrosse/courses/csc321_2017/readings/L15%20Exploding%20and%20Vanishing%20Gradients.pdf,...
  • I. et al. Comparative analysis and forecasting of covid-19 cases in various european countries with arima, narnn and lstm approaches. ELSEVIER,...
  • K. A. et al. Forecasting of covid-19 using deep layer recurrent neural networks (rnns) with gated recurrent units (grus) and long short-term...
  • R. K. et al. Covid-19 in iran: Forecasting pandemic using deep learning. Hindawi, 2021b.
  • M. C. K. A. R. R. J. Filippo Maria Bianchi, Enrico Maiorino. Recurrent Neural Networks for Short-Term Load Forecasting. Springer, first edition,...
  • U. . M. Johns Hopkins. Coronavirus Resource Center . https://coronavirus.jhu.edu/, 2022. [En línea; acceso Enero/2022].
  • M. S. Ke-Lin Du. Neural Networks and Statistical Learning. Springer, 2019.
  • MIT. MIT 6.S191: Recurrent Neural Networks and Transformers. https://youtu.be/QvkQ1B3FBqA, 2022a. [En línea; acceso Enero/2022].
  • MIT. Mit 6.s191: Recurrent neural networks and transformers, mar 2022b. URL https://www.youtube.com/watch?v=QvkQ1B3FBqA&t=2101s&ab_channel=AlexanderAmini....
  • G. Nacional. Datos Abiertos de Colombia. https://www.datos.gov.co/, 2022. [En línea; acceso 6/marzo/2022].
  • S. rekja Hanumanthu. Role of intelligent computing in covid-19 prognosis: A state-of-the-art review. Pre-proof, 2020.
  • Y. Tamura. Simple RNN: the first foothold for understanding LSTM. https://data-science-blog.com/blog/2020/06/17/simple-rnn-the-first-foothold-for-understanding-lstm/,...
  • A. M. S. H.-G. Zimmermann. Recurrent neural networks are universal approximators. International Journal of Neural Systems, 17, 2007.