Global Warming Research Based on GA-BP Neural Network Model

Abstract

Since the industrial revolution, global carbon emissions have continued to increase, and after entering the 21st century, the trend of global warming has become more and more obvious. This paper first use the collected 10-year monthly average temperature data and divide them into two groups: 2012.3~2022.2 and 2022.3~2022.10, then conduct the Mann-Whitney U test, and find that the temperature increase in March 2022 did not increase significantly; then we three models are established to predict the future global temperature, the first model is ARIMA, the second model is BP neural network, and the third model is GA-BP neural network. Three models are used to predict the time when the future temperature will reach 20°C, ARIMA predicts the time is 2050, BP is 2046, and GA-BP is 2045. Finally, this paper use MAE, MAPE, and RMSE to evaluate the accuracy of the model, and find that GA-BP has the highest accuracy, followed by BP, and ARIMA is the worst.