The clustering for graph and network data has a wide application in modern life, such as social networking. However, more challenges crop up along with the needs. High computational cost, sophisticated graphs, and high dimensionality and sparsity are the major concerns. With some special transformations, the issues can be transformed into graph cut issues.
Structural Clustering Algorithm for Network (SCAN) is one of the algorithms that searches for well-connected components in the graph as clusters.
Please take a look at the R codes file ch_06_scan.R from the bundle of R codes for the previously mentioned algorithms. The codes can be tested with the following command:
> source("ch_06_scan.R")
Social network has become the most popular online communication method nowadays. The analysis of SNS becomes important, because of the requirement of security, business, control, and so on.
The foundation of SNS is a graph theory, especially for SNS mining, such as finding social communities and abuse of SNS for bad purpose.
The clustering for SNS is an inherent application to find a community (or community detection). Random walk is another key technology for SNS analysis and is used to find communities.