Population Prediction of China Based on ARIMA-LSTM Combined Model

Abstract

In recent years, the birth rate in China has been declining continuously. Accurate population prediction is of great significance for the government to formulate population macro-adjustment policies. In this paper, the residual optimization method is used to establish the combined prediction model of ARIMA-LSTM. It is found that the precision of combinatorial prediction is better than that of each single prediction, and the combinatorial prediction with residual optimization has better generalization ability. In addition, the forecast results show that China's population will continue to decline in 2023.