A linear time series analysis of carbon price via a complex network approach

Identifying the essential characteristics and forecasting carbon prices is significant in promoting green transformation. This study transforms time series into networks based on China’s pilots by using visibility graph, mining more information structure features. Then, we calculate nodes’ similarity to forecast link prediction. To improve predicted accuracy, notice node distance introduce weight coefficient, measuring impact of different nodes future nodes. Finally, this divides eight communities hierarchical clustering similarities between these pilots. The results show that are “small world” except for Chongqing Shenzhen pilots, all which “scale-free” Shanghai Tianjin Compared with other methods, proposed method has good performance. Moreover, divided three clusters, indicating a higher their price-setting schemes same community. Based analysis provides references trading related enterprises.