Prediction model of urban epidemic prevention and control based on multivariate Bi-A-LSTM

Abstract

The Novel Coronavirus outbreak has since spread rapidly across the globe, seriously affecting the quality of life and economic development of all countries. By predicting the epidemic situation in a certain area, government departments can take corresponding measures to prevent and control the epidemic according to the forecast results. However, with the implementation of various prevention and control measures, vaccination and the impact of virus mutation, the traditional epidemic model and regression model have limitations in the prediction performance, there is a large error in the prediction accuracy. To improve the prediction accuracy, this paper proposes an Attentional mechanism-based LSTM network (A-LSTM), which takes the multiple factors affecting the epidemic trend as the input of the model. Bidirectional A-LSTM is constructed by a-LSTM neural network unit, and the best fitting degree is obtained by training in bidirectional A-LSTM network. Multivariate Bi-A-LSTM epidemic prevention and control prediction model was obtained. In this paper, the actual data as reference, the average absolute error, average absolute percentage error, root mean square error as the evaluation index of the model, and the improved model and other models are compared, experimental results show that the improved model is more accurate than the traditional model in prediction accuracy.