Analysis of influencing factors and carbon emission scenario prediction during building operation stage

Accelerated urbanization driven by the expanding building industry has increased carbon emissions. Emissions during the building operation stage account for more than half of the total carbon emissions associated with the entire house-building process in China, highlighting the substantial potential for energy saving and emission reduction. In this paper, the factors influencing building carbon emission were categorized into demographic, economic, and technological factors based on environmental impact = population × affluence × technology equation. The importance of these factors was ranked using the adaptive boosting (AdaBoost). Second, employing the stochastic impacts by regression on population, affluence, and technology model, we constructed influence factor models for the operation stage. Multiple covariance tests and elimination and regression analyses were conducted to eliminate multicollinearity and identify key factors. Using the scenario analysis method, we developed baseline, low carbon, and high-carbon scenarios to predict future building carbon emissions, estimated the carbon peak time over the next 25 years, and projected carbon emission values for different stages across different scenarios. Third, to further validate prediction accuracy, the AdaBoost algorithm was applied to model future building carbon emissions. Three evaluation metrics—coefficient of determination (R2), root mean square error, and mean absolute error—were used to assess the performance of the prediction model, which demonstrated high accuracy. Finally, based on the findings, countermeasures for achieving low-carbon emissions in buildings have been proposed.