Predicting COVID-19 trends in Canada: a tale of four models

Abstract

This study aims to offer multiple-model informed predictions of COVID-19 in Canada, specifically through the use of deep learning strategy and mathematical prediction models including long-short term memory network, logistic regression model, Gaussian model, and susceptible-infected-removed model. Using the published data as of the 10th of April 2020, the authors forecast that the daily increased number of infective cases in Canada has not reached the peak turning point and will continue to increase. Therefore, Canada's healthcare services need to be ready for the magnitude of this pandemic.

Keywords

• diseases
• logistics
• learning (artificial intelligence)
• regression analysis
• health care
• multiple-model informed predictions
• tale
• covid-19 trends
• canada
• infective cases
• daily increased number
• susceptible-infected-removed model
• gaussian model
• logistic regression model
• long-short term memory network
• mathematical prediction models
• deep learning strategy