561
个编辑
更改
跳到导航
跳到搜索
第211行：
第211行： − 对于矩阵指数函数+ 第221行：
第221行： − +
→Radial-volume centralities exist on a spectrum
for matrix exponentials, where
for matrix exponentials, where
对于矩阵指数函数，
* <math>\beta</math> is a discount parameter which ensures convergence of the sum.
* <math>\beta</math> is a discount parameter which ensures convergence of the sum.
K为步长，A_R是邻接矩阵的转秩，贝达是保证收敛的折扣参数。
Bonacich's family of measures does not transform the adjacency matrix. [[Alpha centrality]] replaces the adjacency matrix with its [[resolvent formalism|resolvent]]. Subgraph centrality replaces the adjacency matrix with its trace. A startling conclusion is that regardless of the initial transformation of the adjacency matrix, all such approaches have common limiting behavior. As <math>\beta</math> approaches zero, the indices converge to [[Centrality#Degree centrality|degree centrality]]. As <math>\beta</math> approaches its maximal value, the indices converge to [[Centrality#Eigenvector centrality|eigenvalue centrality]].<ref name=Benzi2013/>
Bonacich's family of measures does not transform the adjacency matrix. [[Alpha centrality]] replaces the adjacency matrix with its [[resolvent formalism|resolvent]]. Subgraph centrality replaces the adjacency matrix with its trace. A startling conclusion is that regardless of the initial transformation of the adjacency matrix, all such approaches have common limiting behavior. As <math>\beta</math> approaches zero, the indices converge to [[Centrality#Degree centrality|degree centrality]]. As <math>\beta</math> approaches its maximal value, the indices converge to [[Centrality#Eigenvector centrality|eigenvalue centrality]].<ref name=Benzi2013/>