Case Analysis of Epidemic Prediction Based on Fully Convolutional Spatio-temporal Network
This paper expounds using chart learning to capture potential dynamic features based on case data.It proposes a spatiotemporal mixed effects convolutional model to predict the trend of data,encode potential propagation patterns into a learning model,extract the equivalent network of effect propagation from the time series of data,and establish a mixed message transmission mechanism.It evaluates the method on multiple publicly available datasets.The experimental results demonstrate the superiority of the model and demonstrate the practical value of graph neural networks in case prediction.
NETL
NSTL
万方数据
