PREDICTION OF EPIDEMIC CASES NUMBER BASED ON GRAPH NEURAL NETWORK AND TRANSFER LEARNING

Predicting the number of cases of an epidemic is essential for studying epidemiology and ensuring health and safety, but the existing researches work rarely consider factors such as real-time mobility data, which poses a challenge to the prediction of the number of cases. Therefore, based on the graph neural network GNN, a new computing framework-information aggregation network IAN, is proposed, which not only considers the characteristics of regional case data, but also considers the characteristics of population mobility data between various regions. In order to optimize the early prediction model of each country, the transfer learning method TL was added on the basis of this framework. Experimental results on data sets of four European countries show that IAN and IAN-TL are significantly better than traditional methods and can effectively reduce prediction errors.