Machine Learning Approaches for COVID-19 Forecasting

Autor: Nailah Al-Madi, Othman Istaiteh, Tala Owais, Saleh M. Abu-Soud
Rok vydání: 2020
Předmět:
Zdroj: IDSTA
2020 International Conference on Intelligent Data Science Technologies and Applications (IDSTA)
Popis: COVID-19 (Coronavirus) pandemic tends to be one of the most global serious issues in the last century. Furthermore, the world did not face any similar experience regarding the spread of the virus and its economic and political impacts. Forecasting the number of COVID-19 cases in advance could help the decision-makers to take proactive measures and plans. This paper aims to provide a global forecasting tool that predicts the COVID-19 confirmed cases for the next seven days in all over the world. This paper applies four different machine learning algorithms; The autoregressive integrated moving average (ARIMA), artificial neural network(ANN), long-short term memory (LSTM), and convolutional neural network (CNN) to predict the COVID-19 cases in each country for the next seven days. The fine-tuning process of each model is described in this paper and numerical comparisons between the four models are concluded using different evaluation measures; mean absolute error (MAPE), root mean squared logarithmic error (RMSLE) and mean squared logarithmic error (MSLE).
Databáze: OpenAIRE