国际石油价格复杂网络的动力学拓扑性质

为了分析国际石油价格波动的变化特征,依据粗粒化方法,将1986年1月2日至2009年11月17日美国西得克萨斯原油现货逐日离岸价格转化为由3个表示涨落字符{R,e,D}构成的102种5字串组成的价格波动状态为网络节点(即连续5天的价格波动组合),按照时间顺序连边,构建原油价格有向加权波动网络,用复杂网络拓扑结构记录原油价格波动周涨落的信息,计算网络的度与度分布、聚类系数、最短路径长度等动力学统计量.结果表明,石油价格序列的波动网络节点度与累积度分布、节点度值与秩呈幂律分布,前32个节点(模态)的节点度值比较大,且这些石油价格波动模态中都包含表示价格变化上升趋势的字符.原油价格波动网络部分节点的中介中心性能力较强,24.5%的节点承担了网络80.97%的中介中心性功能,平均路径长度2.285,距离是2和3的情况占总数的86.8%,从网络结构拓扑性质这一新的角度验证了原油价格变化的复杂特征.这些对于我们识别具有拓扑统计重要性的节点模态,理解原油价格波动的内在规律和价格变化信息的传导有一定指导意义.

Dynamic analysis on the topological properties of the complex network of international oil prices

To analyze the dynamic of crude oil price data using homogenous partition of coarse graining process,OK WTI Spot Price FOB from January 1986 to November 2009 was transformed into symbolic sequences consisting of three characters (R,e,D). The vertices of the oil price fluctuation network were 5-symbol strings in the number of 102(i.e.102 fluctuation patterns in durations of 5 days) linked in the network's topology by time sequence. It contained integrated information about interconnections and interactions between fluctuation patterns of oil price in network topology. The dynamical statistics of the degree, degree distribution, clustering coefficient and the shortest path length were calculated. The results indicated that the degree of the oil price sequence fluctuation network and the accumulated degree distribution show a power-law distribution, so did the node degree and the rank. The degree of the former 32 nodes was larger, and most oil price fluctuation models had the character with the escalating trend. Some nodes of the oil price fluctuation network had stronger ability in between centrality, and 24.5 percent of the nodes took part in the 80.97 percent of Between Centrality function. The average path length was 2.285, and the paths with length 2—3 were 86.8 percent, which verified the complex characteristics of the oil price fluctuation in network topology. This paper had guiding significance in identifying the topological importance of the node model and understanding the inherent law and information of the oil price fluctuation.