%0 Journal Article
%T 基于趋势性时间序列的全国碳排放量预测研究<br>Research on National Carbon Emission Forecasting Based on Trend Time Series
%A 钟进
%A 李宗航
%J Operations Research and Fuzziology
%P 3870-3881
%@ 2163-1530
%D 2023
%I Hans Publishing
%R 10.12677/ORF.2023.134389
%X 根据最新数据和趋势分析，中国碳排放量一直处于持续增长的形势，尽管中国政府已经采取推广清洁能源、加强能源效率、推动碳交易市场建设等一系列措施以达到节能减排目的，但中国碳排放量仍然面临诸多问题。因此本文基于2019年1月至2023年3月的1551条日频数据，利用SARIMA、LSTM以及SVR等模型综合考虑时间序列的趋势性，对比分析了不同模型对于碳排放数据的预测效果，结果表明SARIMA模型对于碳排放数据的预测效果优于其余两个模型。<br />
According to the latest data and trend analysis, China’s carbon emissions have been in a situation of continuous growth, and although the Chinese government has taken a series of measures to promote clean energy, strengthen energy efficiency, and promote the construction of carbon trading market to achieve energy conservation and emission reduction, China’s carbon emissions still face many problems. Therefore, in this paper, based on 1551 daily frequency data from January 2019 to March 2023, we compare and analyze the prediction effect of different models for carbon emission data using SARIMA, LSTM and SVR models considering the trend of time series, and the results show that SARIMA model has better prediction effect than the remaining two models for carbon emission data.
%K 碳排放，时间序列，LSTM模型，SARIMA模型，SVR模型<br>Carbon Emissions
%K Time Series
%K LSTM Model
%K SARIMA Model
%K SVR Model
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=70776