Forecasting model of COVID-19 pandemic in Malaysia: An application of time series approach using neural network

Abstract

COVID-19 has spread to more than a hundred countries worldwide since the first case reported in late 2019 in Wuhan, China. As one of the countries affected by the spread of COVID-19 cases, the local government of Malaysia has issued several policies to reduce the spread of this outbreak. One of the measures taken by the Malaysian government, namely the Movement Control Order, has been carried out since March 18, 2020. In order to provide precise information to the government so that it can take the appropriate measures, many researchers have attempted to predict and create the model for these cases to identify the number of cases each day and the peak of this pandemic. Therefore, hospitals and health workers can anticipate a surge in COVID-19 patients. In this research, confirmed, recovered, and death cases prediction was performed using the neural network as one of the machine learning methods with high accuracy. The neural network model used is the Multi-Layer Perceptron, Neural Network Auto-Regressive, and Extreme Learning Machine. The three models calculated the average percentage error (APE) values for 7 days and obtained APE values for most cases less than 10%; only 1 case in the last day of one method had an APE value of approximately 11%. Furthermore, based on the best model, then the forecast is made for the next 7 days. In conclusion, this study identified that the MLP model is the best model for 7-step ahead forecasting for confirmed, recovered, and death cases in Malaysia. However, according to the result of testing data, the ELM performs better than the MLP model.

Time Series Forecasting of Covid-19 Confirmed Cases in Turkey With Stacking Ensemble Models

“…Within the scope of this study, many studies have been carried out in order to obtain predictions about the course of the pandemic through various time series models. In accordance with this purpose Turkey (Karcıoğlu et al, 2021;Akay & Akay, 2021;Fidan & Yuksel, 2022), Greece (Katris, 2021), Portugal (De Oliveira, et al, 2021, Italia (Ding et al, 2020;Ceylan, 2020;Dehesh et al, 2020), Worldwide (Sevli & Gülsoy;, Maleki et al, 2020, Petropoulos et al, 2022, USA (De Oliveira, et al, 2021;Özen et al, 2021;Shastri et al, 2020;Zeroual et al, 2020), Malaysia (Purwandari et al, 2022), Spain (Ceylan, 2020;Zeroual et al, 2020), Brazil (De Oliveira, et al, 2021Dairi, 2021), İndia (Shastri et al, 2020;Arora et al, 2020;Tandon et al, 2020), Nigeria (Abdulmajeed et al, 2020), France (Zeroual et al, 2020;Ceylan, 2020;Dairi, 2021), Canada (Chimmula & Zhang, 2020), South Korea (Pontoh et al, 2020;Dehesh et al, 2020), China (Dehesh et al, 2020;Zeroual et al, 2020), Iran (Dehesh et al, 2020;Talkhi et al, 2021), Austuralia (Zeroual et al, 2020), Russia (Dairi, 2021), Pakistan (Ali et al, 2020) etc. many studies with COVID-19 time series were conducted for many countries.…”

“…While the scope of these studies mostly consisted of the daily number of positive cases (Kumar & Susan, 2020;Papastefanopoulos et al, 2020;Zeroual et al, 2020;Chimmula & Zhang, 2020;Arora et al, 2020;Ding et al, 2020;Karcıoğlu et al, 2021;Dairi, 2021;Sevli & Gülsoy, 2020;Özen et al, 2021;Tandon et al, 2020;Purwandari et al, 2022;Pontoh et al, 2020;De Oliveira, et al, 2021;Shastri et al, 2020;Talkhi et al, 2021, Petropoulos et al, 2022, which is also the subject of this study, the authors also took into account the daily number of deaths (Karcıoğlu et al, 2021;Sevli & Gülsoy, 2020;Purwandari et al, 2022;Petropoulos et al, 2022, Pontoh et al, 2020Shastri et al, 2020;Talkhi et al, 2021) and the daily recoveries (Karcıoğlu et al, 2021;Sevli & Gülsoy, 2020;Purwandari et al, 2022;Purwandari et al, 2022;Pontoh et al, 2020;Zeroual et al, 2020;Dairi, 2021). In addition, many statistical and machine learning based time series models such as ARIMA (Ali et al, 2020;Tandon et al, 2020;Karcıoğlu et al, 2021;Ding et al, 2020;Özen et al, 2021;Abdulmajeed et al, 2020;Ceylan, 2020;Dehesh et al, 2020;<...…”

“…In this study, it is aimed to evaluate the findings together with the existing studies by preferring the frequently used performance metrics in the time series studies carried out for the COVID-19 pandemic within the scope of the literature. For this purpose, Mean Absolute Error (MAE) in Equation 10 (Ceylan, 2020;Pontoh et al, 2020;Özen et al, 2021;Purwandari et al, 2022;Talkhi et al, 2021), Mean Absolute Percentage Error (MAPE) in Equation 11 (Sevli & Gülsoy, 2020;Karcıoğlu et al, 2021;Özen et al, 2021;Shastri et al, 2020;Arora et al, 2020;Talkhi et al, 2021) and Root Mean Square Error (RMSE) in Equation 12 (Ceylan, 2020;Pontoh et al, 2020;Özen et al, 2021;Talkhi et al, 2021;Purwandari et al, 2022) were used to compare model performances as evaluation metrics. Models with minimum values are selected in the most model selection according to the criteria below.…”

Attributable Fraction and Forecasting for COVID-19 Confirmed Cases in Nigeria Using Facebook- Prophet Machine Learning Model

“…The authors found some underestimation and overestimation of the daily cases. [5,17,21,23] The authors applied deep learning methods, these methods can understand trends, dependencies, and structures.…”

Comparison of Parametric and Nonparametric Forecasting Methods for Daily COVID-19 Cases in Malaysia

“…The total confirmed cases are predicted using the number of recovery patients. Purwandari et al [33] forecasted COVID-19 cases using the MLP, Neural Network Autoregressive (NNA), and Extreme Learning Machine (ELM). All methods achieved less than 12% MAPE for the days forecast.…”