Application of Generalized Regression Neural Network and Sarima Model in Prediction of AIDS

Abstract

Generalized regression neural network (GRNN) is highly fault-tolerant and robust, which is suitable to solve nonlinear problems, and is currently widely used in prediction research. The seasonal autoregressive integrated moving mean model (Sarima) captures the seasonal, periodicity of historical data very well. In order to arouse people's concern about health and keep away from AIDS again, in the study, we applied GRNN and Sarima model to explore the prediction of the incidence and mortality of AIDS Based on historical AIDS data in Guangxi. We established the high-precision Sarima (2,0,1)(1,0,1)12 model and GRNN network with spread 0.4 to predict the numbers of monthly AIDS deaths and reported cases from November 2019 to December 2021. The results of the prediction analysis indicated that if the prevention and control efforts will not be increased, the incidence of AIDS in Guangxi may remain high, and the number of AIDS deaths per month may show an upward trend, which provides early warning and scientific reference for the prevention departments to optimize the allocation of resources in advance.