A method to extend existing document clustering procedures in order to include relational information
Tijn Witsenburg, Hendrik Blockeel · Lirias · 2008
We consider the problem of clustering nodes in a graph, where each node has also internal content (e.g., the Web, where nodes are web pages). In this context we can distinguish two kinds of information: content information and structural information. Standard clustering methods use content information only, while graph clustering methods are usually based on the graph structure. Relatively recently, researchers have proposed to combine both types of information. In this paper we propose a very simple, yet hitherto unexplored, method for doing this by extending existing clustering procedures that use content information. 1.