Spatiotemporal prediction of carbon emissions using a hybrid deep learning model considering temporal and spatial correlations

Task This paper presents a deep learning-based hybrid prediction model for carbon emissions. The model enables the prediction of future carbon emissions, in single- and multi-step scenarios, by using historical time series data. The monthly ODIAC data for the three major urban agglomerations in China, namely, the Yangtze River Delta, Pearl River Delta and Beijing–Tianjin–Hebei, were utilised for performance evaluation.

Environment CPU: i5-11400H @ 2.70 GHz, GPU: NVIDIA GeForce RTX3050 4 GB RAM: 16 G. Encoding language：Python 3.7.12

Dataset The ODIAC datasets used in this study are available at http://www.odiac.org/index.html Preprocessed data： Yangtze River Delta： GYH62.xlsx is the preprocessed feature matrix LJJZ.xlsx is the adjacency matrix JLJZ.xlsx is the distance weight matrix Pearl River Delta： GYH62.xlsx is the preprocessed feature matrix LJJZ.xlsx is the adjacency matrix JLJZ.xlsx is the distance weight matrix Beijing-Tianjin-Hebei： GYH62.xlsx is the preprocessed feature matrix LJJZ.xlsx is the adjacency matrix JLJZ.xlsx is the distance weight matrix Example of multi-step prediction：prediction4.xlsx

Model GRU-GCN.py