Reconstructing time series into a complex network to assess the evolution dynamics of the correlations among energy prices

Reconstructing a time series into a complex network can help uncover the dynamic information hidden in the time series. Previous studies mainly focused on the long-term relationship between two energy prices, and traditional econometric methods poorly reflect the evolution of correlations among variables from a short-term perspective. Thus, first, we divide natural gas, coal and crude oil price time series into a series of segments via a set of temporal sliding windows and then calculate the correlation coefficients for each pair of energy prices in each segment. Second, we define the correlation modes based on the correlation coefficients and a coarse graining process. Third, we reconstruct the time series into a complex network to assess the evolution dynamics of the correlations among energy prices. The results show that a few major correlation modes and transmission patterns play important roles in the evolution. The evolution of the correlation modes among energy prices exhibits a significant cluster effect. Approximately 30 days is a turning point at which one type of cluster transforms into another type. Then, we improve the betweenness centrality algorithm to measure the media capability of the correlation mode in the evolution process of different clusters. Based on the transmission probabilities between clusters, we can determine the evolution direction of the correlation modes based on energy prices. These results are useful for monitoring fluctuations in energy prices and making decisions for risk avoidance.