A hybrid method for constructing High Level Architecture of BBS user network

It is useful to understand the High Level Architecture (HLA) of the user network of Bulletin Board Systems (BBS) for some applications. In this paper, we construct the HLA of a BBS user network through hybrid static and dynamic analysis of the quantitative temporal graphs that are extracted from the BBS entries. We detect the HLA framework first though the static structural analysis of the aggregation of the temporal graphs. Then, we identify the HLA components including communities, community cores, and hubs elaborately through the dynamic analysis of the quantitative temporal attributes of nodes. The hybrid method guarantees the HLA quality as it removes the false components from the HLA. It controls the computational cost at a low level also. A metric is proposed to evaluate the HLA efficiency in information transmission. The experiments show that the HLA constructed by the hybrid method outperforms that constructed by the comparative method.