Carbon futures price forecasting based with ARIMA-CNN-LSTM model

Abstract

In this paper, we introduced an ARIMA-CNN-LSTM model to forecast the carbon futures price. The ARIMA-CNN-LSTM model employs the ARIMA model and the deep neural network structure that combines the CNN and LSTM layers to capture linear and nonlinear data features. In ARIMA-CNN-LSTM model structure, the ARIMA is used to capture the linear features. The Convolutional Neural Network (CNN) is used to capture the hierarchical data structure while the Long Short Term Memory network (LSTM) is used to capture the long-term dependencies in the data. Comprehensive performance evaluation has been conducted using weekly carbon futures price. Results have confirmed that ARIMA-CNN-LSTM model can achieve better prediction accuracy than the benchmark models, in terms of Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE) performance measures.