CAJ | 학술논문

系统分析了中国CO2排放的影响因素，在此基础上建立了基于高斯过程回归的CO2排放预测模型，运用历史数据进行模型精度检验，并与传统的GM （1，1）模型、人工神经网络和支持向量机的预测结果比较。结合情景设计，预测了中国“十三五”时期的CO2排放量和CO2排放强度。结果表明：高斯过程回归模型具有显著的精度优势，中国能达到2020年CO2排放强度较2005年下降40％～45％的减排目标。指出：对于CO2减排，应结合各地区的实际情况灵活处理，以调整产业结构、优化能源结构、推动技术创新为重点，不可片面牺牲经济发展。
계통분석료중국CO2배방적영향인소，재차기출상건립료기우고사과정회귀적CO2배방예측모형，운용역사수거진행모형정도검험，병여전통적GM （1，1）모형、인공신경망락화지지향량궤적예측결과비교。결합정경설계，예측료중국“십삼오”시기적CO2배방량화CO2배방강도。결과표명：고사과정회귀모형구유현저적정도우세，중국능체도2020년CO2배방강도교2005년하강40％～45％적감배목표。지출：대우CO2감배，응결합각지구적실제정황령활처리，이조정산업결구、우화능원결구、추동기술창신위중점，불가편면희생경제발전。
This paper systematically analyzes the main factors influencing CO2 emission ,and build a multivariate prediction model of CO2 based on the Gaussian process regression .And it utilizes the historical data for the accuracy test ,and compares it with traditional GM （1 ,1） ,artificial neural network and support vector machine prediction model .Then it makes a scenario design to estimate the CO2 emission and emission intensi‐ty during the Thirteenth Five‐year Development Planning .The result shows that the prediction accuracy of GPR model has obvious advantages , and China could achieve the reduction target of decreasing CO2 emission intensity by 40% ‐45% in 2020 compared with 2005 .Finally ,it proposes that different regions should implement CO2 emission according to own actual situation ,and focus on adjusting industrial structure ,optimizing energy structure and promoting technological advancement emissions ,rather than rely on sacrificing economic growth .