Abstract:
CORONAVIRUS DISEASE (COVID-19) is the infectious disease caused by
the coronavirus that was first discovered in Wuhan City, China, and then spread
throughout the world. Many researchers have proposed various methods to predict the
spread of viruses. Predicting the number of COVID-19 patients is a crucial task in the
effort to assist governments and healthcare departments respond rapidly to outbreaks.
One type of prediction method is Artificial Neural Network (ANN), which is much
more flexible and can handle more complicated and unassuming cases than the
regression method. There are many ANN algorithms. Among them, the
backpropagation algorithm is used in the proposed system. The backpropagation
algorithm is a method for training multilayer feed-forward neural networks. It can be
used to solve predictive problems with good results. Firstly, the proposed system
implements a prediction model to estimate the number of COVID-19 sufferers in
ASEAN Countries using a backpropagation neural network with Gradient Descent
and a Backpropagation neural network with Stochastics Gradient Descent Optimizers.
Among them, the method that produced the best performance is used to predict the
future number of COVID-19 cases. And then these predicted results are used to
decide the risk category of a country with Fuzzy Inference System. To evaluate the
performance of the prediction methods for the number of COVID-19 sufferers, Root
Mean Square Error (RMSE) is used and compared. According to the experimental
results, the Backpropagation neural network with Stochastics Gradient Descent
method has a better performance than the Backpropagation neural network with
Gradient Descent method. The accuracy of the Fuzzy Inference method for the
classification of the risk category of each country is calculated many times by using
the preexisting actual trend data. As a result, the proposed system can be useful for
risk categorization and long-term outbreak prediction in epidemics like COVID-19.
