Modelos híbridos SARIMA-ANN para pronósticos de la COVID-19 en el Perú
DOI:
https://doi.org/10.21754/iecos.v22i1.1332Palabras clave:
Modelos ARIMA, Redes Neuronales Autoregresivas, Perceptron Multicapas, Modelos híbridos NNAR-ARIMA, Modelos híbridos MLP-ARIMAResumen
Se ha construido modelos híbridos ANN-ARIMA por remodelamiento, para realizar los pronósticos de los nuevos casos de contagios por Covid-19 en el Perú, para ello se extrajo y uso los casos confirmados de Covid-19 entre el periodo 06/03/20 hasta el 28/02/21, desde la plataforma de los datos abiertos del Ministerio de Salud. Los resultados hallados indican que los 02 mejores modelos corresponden al modelo hibrido multiplicativo NNAR (27,1,6) * ARIMA(3,0,2)(1,0,1), y al modelo hibrido aditivo NNAR (27,1,6) + ARIMA(1,0,1), cuyos valores del error medio absoluto porcentual(MAPE) se diferencian en tan solo el 0.575% por lo que proporcionan casi los mismos pronósticos. Considerando el promedio de los valores del MAPE para los 03 mejores modelos de cada categoría de modelamiento se ha determinado que los modelos híbridos NNAR-ARIMA son mejores que los modelos híbridos MLP-ARIMA, que modelos híbridos aditivos NNAR+ARIMA tienen una superioridad del 1.20% sobre los modelos híbridos multiplicativos NNAR*ARIMA; mientras que la superioridad del modelo hibrido aditivo MLP+ARIMA sobre el modelo hibrido multiplicativo MLP*ARIMA alcanza al 2.31%.
Descargas
Citas
BROCKWELL, P. & DAVIS R. (1991) Time Series: Theory and Methods. Colorado State University, second edition, Springer-Verlag, New York Inc.
BOX G.E.P & JENKINS G.M (1970) Time Series Analysis: Forecasting and Control. San Francisco: Holden-Day.
CEYLAN Z. (2020) Estimation of COVID-19 prevalence in Italy, Spain, and France. Science of the Total Environment 729 (2020) 138817.
CYBENKO G. (1989) Approximation by superpositions of a sigmoid function. Mathematics of Control Signals and Systems 2, 303–314.
DEHESH T; FARD H.A. M.; & DEHESH P. (2020) Forecasting of COVID-19 Confirmed Cases in Different Countries with ARIMA Models. MedRxiv preprint,
DOI: https://doi.org/10.1101/2020.03.13.200 35345.
DING G. LI X.; JIAO F.; & SHEN Y. (2020) Brief Analysis of the ARIMA model on the Covid 19 in Italy. MedRxiv preprint https://doi.org/10.1101/2020.04.08.20058636
GANINY S. & NISAR O. (2021) Mathematical modeling and a month ahead forecast of the coronavirus disease 2019 (COVID‑19) pandemic: an Indian
Scenario. Modeling Earth Systems and Environment. Modeling Earth Systems and Environment. 2021 Jan: 1-12. DOI: 10.1007/s40808-020-01080-6.
GOBIERNO DEL PERU (2021) Datos abiertos Covid-19 Plataforma Nacional Ministerio de Salud-MINSA. https://www.datosabiertos.gob.pe/dataset/casos-positivos-por-covid-19-ministerio-de-salud-minsa
HAMILTON, J. D. (1994) Time Series Analysis. Princeton University Press, Princeton NJ.
HYDMAN R.J. and ATHANASOPOULOS G. (2013) Forecasting: Principles and Practice. Monash University, Australia.
HORNI.K.; STINCHCOMBE, M.; & WHITE H. (1989) Multilayer feed forward networks are universal approximators. Neural Networks 2, 359–366.
HOOKER R. H. (1901) “On the correlation of the marriage-rate with trade." Journal Roy. Stat. Soc., London, vol. 64, p. 485, 1901.
HORNIK K.; STINCHCOMBE M.; & WHITE H. (1990) Universal approximation of an unknown mapping and its derivatives using multilayer feed forward networks. Neural Networks 3, 551–560.
NIELSEN A. (2020) Practical Time Series Analysis. Published by O’Reilly Media, Inc., 1005 Gravenstein Highway North, and Sebastopol. USA
PANKRATZ A. (1983) Forecasting with Univariate Box‐Jenkins Models: Concepts and Cases First edition. John Wiley & Sons, Inc.
PERONE G. (2020a) An ARIMA model to forecast the spread and the final size of Covid 19 epidemic in Italy. No. 20/07. HEDG, c/o Department of Economics, University of York,
PERONE G. (2020b) Comparison of ARIMA, ETS, NNAR and hybrid models to forecast the second wave of COVID-19 hospitalizations in Italy. October 2020 https://arxiv.org/ftp/arxiv/papers/2010/2010.11617.pdf
SAFI S.K. & SANUSI I.S. (2021) A hybrid of artificial neural network, exponential smoothing, and ARIMA models for COVID-19 time series forecasting. Model Assisted. Statistics and Applications 16 (2021) 25–35 25 DOI 10.3233/MAS-210512 IOS Press.
SHIBATA K. & IKEDA Y.(2009) “Effect of number of hidden neurons on learning in large-scale layered neural networks,” in Proceedings of the ICROS-SICE International Joint Conference 2009 (ICCASSICE ’09), pp. 5008–5013, August 2009.
TRENN S. (2008) “Multilayer perceptrons: approximation order and necessary number of hidden units,” IEEE Transactions on Neural Networks, vol. 19, no. 5, pp. 836–844.
URIEL J.E.(1985) Análisis de series temporales: Modelos ARIMA. Paraninfo Madrid España
WANG L.; ZOU H.; SU J.; LI L.; & CHAUDHRY S. (2013) An ARIMA-ANN Hybrid Model for Time Series Forecasting. Systems Research and Behavioral Science Syst. Res. 30, 244–259 (2013). Research paper.
WEI W.W.S (1990) Time series analysis univariate and multivariate methods. Temple University. First edition Addison-Wesley, Reading, MA.
YULE G.U. (1909) The Applications of the Method of Correlation to Social and Economic Statistics. Journal of the Royal Statistical Society, Vol. 72, No. 4 (Dec., 1909), pp. 721-730
YULE G.U (1927) On a method of investigations periodicities in disturbed series, with special reference to Wolfer’s sunspot numbers. Philos. Trans. Roy. Soc. London Ser. A 226 267-298.
ZHANG G.P. (1998) Linear and nonlinear time series forecasting with artificial neural networks. Ph.D. Dissertation, Kent State University, Kent, OH.
ZHANG G. P.(2003). Time series forecasting using a hybrid ARIMA and neural network model. Neurocomputing, 50: 159-75.v
ZHANG G.P. & QI M. (2005) Neural network forecasting for seasonal and trend time series. European Journal of Operational Research, 160, pp. 501-514.
Descargas
Publicado
Cómo citar
Número
Sección
Licencia
Derechos de autor 2021 Alipio Francisco Ordoñez Mercado
Esta obra está bajo una licencia internacional Creative Commons Atribución 4.0.
CC BY 4.0 DEED Attribution 4.0 International
