Effectiveness Estimation of Interventions against COVID-19 Using AI

Abstract

The interventions implemented in each region to prevent the spread of COVID-19 depend on a great many factors. In order to estimate the effects of the interventions, a large number of those factors need to be analyzed in combination. Fujitsu Laboratories have developed Wide Learning which can search for combinations of data items exhaustively and rapidly to find out important combinations. Using this Wide Learning, we analyzed the success or failure of the interventions in each region through August 2020. Furthermore, based on the results, we use a what-if analysis to estimate the effect of each intervention in each region and at each point in time.