Research on Carbon Emission Prediction Model for Building Structure Construction Based on ARIMA-BPNN

In this study, a hybrid prediction model based on the fusion of ARIMA and BP neural network is proposed to accurately predict the carbon emissions during the construction phase. The model adopts a decomposition-integration strategy, extracting the linear features in the time series by ARIMA model and capturing the nonlinear patterns of the residuals by BP neural network. The empirical analysis is based on the 12-month carbon emission data of a construction project, and the results show that the hybrid model has a root mean square error (RMSE) of 0.0115 on the test set, and 92.18% of the predictions are within 1% of the relative error, which is significantly better than the single prediction model.