The visualization of relatedness

The relationships between documents are usually derived by accounting on the links between them. These linkages may take the form of citation, cocitation, hyperlink, or co-terms. The relatedness between documents is derived from link analysis, which is widely used in the information science disciplines. We use kernel functions to overcome the limitations of the relatedness measurement based on cocitation count which does not account on the indirect relationships. We then elicit all maximal cliques from the resultant kernel graph and visually explore the graph of connected cliques. The meaning of a clique in the graph is interpreted based on a preliminary content analysis of the literatures underlying the clique. We also study overlapping cliques by examining their sharing nodes. The common nodes of adjacent cliques seem to play a bridging role between distinct fields.