基于GM-ARIMA耦合模型的碳排放动态预测与绿色低碳转型路径优化研究——以山东省的碳排放量为例进行预测分析

Research on Carbon Emission Dynamic Prediction and Green and Low-Carbon Transformation Path Optimization Based on GM-ARIMA Coupling Model—Conducting a Predictive Analysis Based on the Carbon Emissions of Shandong Province

摘要: 本文聚焦于农业生产材料投入能力与碳排放的研究，以山东省为例，对山东省碳排放量进行预测分析。首先，建立了碳排放量的灰色时间序列预测模型，并使用粒子群优化算法(PSO)，对该模型的灰作用量、发展系数两个参数进行优化。接着，利用自回归积分滑动平均算法(ARIMA)分析预测模型的残差，并通过网格搜索法优化ARIMA模型的p、d、q参数。最后，利用ARIMA模型的残差预测结果对灰色预测模型进行补偿。本文创新性地提出了灰色理论–自回归滑动平均模型复合模型(GM-ARIMA)，旨在充分发挥灰色理论在处理不确定性和不完全信息方面的独特优势，同时结合ARIMA模型在时间序列分析中的强大预测能力来消除模型残差，从而为农业碳排放量的预测提供一个可靠的基础模型。

Abstract: This article focuses on the research of agricultural production materials input capacity and carbon emissions, using Shandong Province as a case study to predict and analyze the carbon emission levels in the region. Firstly, a gray time series prediction model for carbon emissions was established, and two parameters of the model—gray influence amount and development coefficient—were optimized using the Particle Swarm Optimization (PSO) algorithm. Subsequently, the Autoregressive Integrated Moving Average (ARIMA) algorithm was employed to analyze the residuals of the prediction model, followed by parameter optimization of the ARIMA model through grid search method. Finally, the residual predictions from the ARIMA model were used to compensate for the gray prediction model. This article innovatively proposes a composite model of Gray Theory and Autoregressive Moving Average (GM-ARIMA), aiming to fully leverage the unique advantages of gray theory in handling uncertainty and incomplete information, while integrating the powerful predictive capability of the ARIMA model in time series analysis to eliminate model residuals, thereby providing a reliable foundational model for predicting agricultural carbon emissions.