基于Boosting-ARMA的碳价预测

自回归滑动平均(ARMA)模型是最流行的预测模型之一,而模型选择却是使用ARMA进行预测的难点,尤其是当真实模型的阶数较高时,因此提出Boosting-ARMA预测算法,利用Boosting算法进行最优子集ARMA寻找,自动且高效地完成ARMA模型的识别。模拟实验显示,Boosting-ARMA优于其他方法,用新算法预测碳价实证分析发现,Boosting-ARMA算法可以获得较高的碳价预测准确性并且方便快捷。

Forecasting of Carbon Price Based on Boosting-ARMA Model

Autoregressive moving average (ARMA) model is one of the most popular models in forecasting, but model selection is a difficult point of ARMA modeling, especially when the orders of true model are high.Hence, this paper proposes Boosting-ARMA forecasting algorithm, this algorithm identifies the optimal subset ARMA model via Boosting, finish the identification automatically and efficiently.Simulation results show that Boosting-ARMA is better than other methods.At last, we forecast the carbon price with new algorithm;the empirical results find that Boosting-ARMA can get high forecasting accuracy for carbon price conveniently and efficiently.