Study of temperature change based on ARIMA forecasting and correlation analysis

In this paper, through Mann-Kenddall mutation test method, it was found that the mutation value in the tropical region is 1939, the mutation value in the southern temperate region is 1959, and the mutation value in the northern temperate region is 1938. In this paper, ARIMA time series prediction model and BP neural network model based on genetic algorithm are developed to predict the global future temperature level. We found that the ARIMA time series prediction model is more accurate. Finally, we verified that global temperature is positively correlated with carbon dioxide concentration and population size, and negatively correlated with global forest area through Spearman correlation analysis. This study develops a powerful prediction method that fully exploits the application of intelligent systems in a digital society. The results of the study contribute to the mitigation of global warming as a social problem.