英文摘要 |
Social network analysis is an important tool to observe the formations of online networks and to explore the changes of virtual communities. Amongst the various methods of social network analyses, block modeling is a key approach to analyzing network structures. It can be applied not only to position and role analyses of the individual nodes, but also to cohesive subgroup discoveries of the whole network. Conventional block modeling, operating on the adjacency matrix of the network, tries to adjust and partition the adjacency matrix into blocks that contain all 0s or all 1s approximately, and to investigate the network structure between the approximate blocks. In this research, we apply boolean matrix operation on the adjacency matrix to generate the reachability matrix as the extended target of block modeling. And we use a blogroll network to show the possible applications of the extended block modeling. Via reachability matrices of different distances, we can analyze the network structures based on a variety of loosely cohesive subgroups. Via the longest reachability matrix, we can fit the network perfectly into the bowtie structure, and the fit is also corresponding to the symmetric-acyclic structure. |