Modelling lock-down strictness for COVID-19 pandemic in ASEAN countries by using hybrid ARIMA-SVR and hybrid SEIR-ANN

• Main Authors: Maharani A., Bakar, Norizan, Mohamed, Danang A., Pratama, M. Fawwaz A., Yusran, Nor Azlida, Aleng, Yanuar Z., Arief, L., Niken
• Format: Article
• Language: English
• Published: Taylor & Francis 2021
• Subjects: QA76 Computer software; T Technology (General)
• Online Access: http://ir.unimas.my/id/eprint/35349/2/Modelling%20lock%20down%20strictness%20for%20COVID%2019%20pandemic%20in%20ASEAN%20countries%20by%20using%20hybrid%20ARIMA%20SVR%20and%20hybrid%20SEIR%20ANN.pdf; http://ir.unimas.my/id/eprint/35349/; https://www.tandfonline.com/doi/full/10.1080/25765299.2021.1902606; https://doi.org/10.1080/25765299.2021.1902606
• Institution: Universiti Malaysia Sarawak

ASEAN, include Indonesia, Malaysia, Philippines, Singapore, and Thailand, are the countries with ongoing transmission of SARS-COV-2, the virus that causes COVID-19. The confirmed cases in Indonesia and Philippines are the highest ranks among other ASEAN countries such as Malaysia, Thailand, and Singapore. To reduce the spread of the pandemic COVID-19, each country has implemented the lock-down policy differently, depending on its economic situation. Therefore, the study of the impact of lock-down across the world, particularly in ASEAN countries, is still relevant to do. In this study, we developed the lock-down model in ASEAN countries by using hybrid ARIMA-SVR and hybrid SEIR-ANN. The first hybrid is based on the time series model ARIMA, with the revision of the error is by using SVR. The second hybrid is based on the classical model of infectious diseases, SEIR, which we revise on the prediction part by using ANN. The hybrid is intended to revise the individual prediction model. The data collected per country was started from January 20, 2020 to August 5, 2020. The periods of lock-down in this study are divided into three, namely no lock-down, implemented lock-down, and the new normal periods. The strictness levels of lock-down were predicted for 60 days ahead. The results showed that the hybrid ARIMA-SVR had smaller RMSE compared with individual ARIMA, similarly, hybrid SEIR-ANN predicted S, E, I, and R more accurately compared with individual SEIR model. It has been also found that the lock-down was most effectively implemented in Thailand, Singapore, and Malaysia, whereas Indonesia and Philippines were inefficient countries to enforce the restriction. It is indicated by the number of cases increased significantly during the restriction periods in both countries.