Analysis and prediction of carbon emissions from public building operation based on stacking ensemble strategy: the case of Xi 'an

Global climate change is a serious problem facing all countries at present, and CO2 among greenhouse gases is considered to be the main contributor. Public buildings (PBs) are representative of high carbon-emitting buildings, which have great potential for energy conservation and carbon reduction. The analysis and accurate prediction of carbon emissions (CEs) in the operation stage of PBs are important for formulating reasonable and feasible carbon emission reduction pathways and realizing sustainable development. Firstly, this paper updates the electricity carbon emission factor (ECEF) and heating carbon emission factor (HCEF) in Shaanxi Province for 2018–2022, so as to improve the accuracy of CEs calculation. Secondly, the influencing factors of CEs in the operation stage of PBs are comprehensively considered, and the RFS method is proposed to analyze the importance and relevance of each variable to CEs. After that, a CE prediction model based on a stacking ensemble strategy is proposed to predict daily CEs during the operation stage of PBs. Finally, typical operational data from PBs of different types in Xi'an, Shaanxi Province, China are obtained for model validation, and the application of the model is discussed. The results show that the model can accurately estimate and predict the CEs of PBs during operation. The research results provide methodological support for relevant departments to use multi-source data to forecast CEs, and provide a practical and effective reference for decision-makers to reduce CEs.