利用ARIMA-SVM模型的碳排放交易价格预测

Carbon emissions trading price prediction using the ARIMA-SVM model

为了帮助企业、投资者和市场监管部门优化碳排放市场参与行为,需要对碳排放交易价格进行合理有效的预测。考虑到碳排放交易价格时间序列同时具有线性和非线性2种特征,选择ARIMA-SVM融合模型运用到碳排放交易价格预测中,发挥该模型预测精度高的优势。运用ARIMA-SVM模型、ARIMA模型、SVM模型和Db6-SVM模型对湖北碳排放交易价格进行8期预测。通过4种模型预测值的MSE值和MAE值确定预测精度,对比预测精度,探究ARIMA-SVM模型是否为准确有效的预测模型,实证结果表明:ARIMA-SVM模型的MSE值为0.1770,是4种模型的最低值;MAE值为0.3387,是4种模型的次低值。可以认为ARIMA-SVM模型的预测精度最高,是一种有效的且精度高的碳排放价格预测模型,可用于碳排放交易价格预测,可以为碳排放交易参与企业和各方投资者把握价格波动趋势,增强防范能力提供保障,也可以为市场监管职能部门防止碳排放交易价格过度波动及时制定有效措施。

Accurate and effective carbon emissions trading price prediction is of great significance to companies,for optimizing behavior of investors and market regulators involved in carbon emissions trading.The Autoregressive Integrated Moving Average model(ARIMA)-Support Vector Machine(SVM)model with a consideration of linear and non-linear features of the carbon emissions trading price time series is applied to the carbon emissions trading price prediction,and the model has the advantage of high prediction accuracy.The ARIMA-SVM model,ARIMA model,SVM model and Db6-SVM model are used to make 8 stage forecast of Hubei carbon emissions trading prices.The prediction accuracy is determined by the MSE and MAE values of the predicted values of the four models,and the prediction accuracy is compared in order to explore whether the ARIMA-SVM model is an accurate and effective prediction model.The empirical results show that the MSE value of the ARIMA-SVM model is 0.1770,which is the lowest value of the four models;the MAE value is 0.3387,which is the second lowest value of the four models.It can be considered that the ARIMA-SVM model has the highest prediction accuracy,and is an effective and high-precision carbon emission price prediction model.It can be used for carbon emission transaction price prediction,and can grasp the price fluctuation trend for companies and investors of carbon emission transactions.In order to enhance the prevention capability and provide protection,it can also formulate effective measures for the market supervision function to prevent excessive fluctuations in carbon emissions trading prices.