[China's Inter-provincial Energy Consumption Emission Reduction Path Based on SNA and CNN-BiLSTM-attention Combined Model].

Carbon emissions between regions exhibit complex spatial correlations， and the achievement of "dual carbon" goals depends not only on provincial factors but also on spatial linkages. Using the carbon emission coefficient method， this study calculates the energy consumption carbon emissions of Chinese provinces from 2000 to 2021 and constructs a spatial correlation network of provincial energy consumption carbon emissions in China. From the perspective of spatial correlation， the GDIM decomposition is applied to analyze the driving factors of energy consumption carbon emissions for the main beneficiary， net beneficiary， broker， and net spillover sectors in the network. A CNN-BiLSTM-attention combined model is used to simulate and predict energy consumption carbon emissions for each sector from 2022 to 2050 under four scenario modes. Emission reduction pathways are proposed based on indicators of the functional role of each sector in the network. The results showed that： ① During the observation period， various sectors underwent different degrees of restructuring， exhibiting significant "gradient transfer" characteristics. ② Economic scale and energy consumption scale were the main positive driving factors for energy consumption carbon emissions at different development stages of each sector， with the largest cumulative contribution. The inhibitory effect of technological progress is gradually becoming apparent. ③ Each sector could basically achieve peak emissions before 2030. To this end， the main beneficiary， net beneficiary， broker， and net spillover sectors should prioritize green innovation， high-speed coordination， high-speed leadership， and green transformation as their primary emission reduction pathways， respectively.