【摘要】 产业关联网络高密度、加权有向和带有自环的特性不同于一般复杂网络,使得传统的基于最短路径的紧密中心性、介数中心性和特征向量中心性算法无法适用于产业关联网络。在随机游走的基础上设计的随机游走中心性算法,有效地解决了上述问题。通过对中国2010年投入产出延长表的计算验证了该算法能有效合理的对产业的重要性进行区分。更多还原

【Abstract】 The importance of the industries in the industry interrelated network can be sorted by complex network centrality. The industry interrelated network is differ from general complex network in three aspects:high-density,weighted-directed and having self-loop,which makes the traditional "shortest path"based centrality algorithms such as betweenness centrality,closeness centrality and eigenvector centrality not applicable to the industry interrelated network. In order to solve the problems above,we design the random walk centrality on the basis of random walk process in the complex network. The random walk centrality has been applied to the Prolong of 2005 Input-Output Table in China and the result shows that the new algorithm is better to sort the importance of the industries than the traditional algorithms。更多还原