Abstract
Links between Web pages can be used as a source of information with which to identify types of Web based communities for which interlinking is common, and to investigate community-based linking practices. These communities could be of individuals, organizations or even information sources. Hyperlink analysis, also known as link analysis, provides a set of techniques to aid data gathering, filtering and processing, as well as providing methods to help interpret results. Hyperlink analysis originated in citation analysis because of the structural similarity between links and citations (Ingwersen, 1998). Whereas links connect Web pages, citations typically connect academic journal articles. Citations have been analyzed for many years by researchers in order to analyze research communities in the scholarly literature (Borgman & Furner, 2002) amongst other things, and the first people to develop link analysis drew from citation analysis experience. Early contributors included Rousseau (1997), who dubbed links sitations, and Ingwersen (1998), who used citation counts to estimate the impact of collections of Web pages, using the assumption that more useful pages would tend to receive more links. The creators of Google also harnessed links to deliver more useful search results based upon a similar assumption (Brin & Page, 1998). One of the earliest academic applications of link analysis was to identify “communities” of information around a particular topic (Larson, 1996). Since then many others have used links to help identify Web communities (Flake, Lawrence, Giles, & Coetzee, 2002), or to describe them (Foot, Schneider, Dougherty, Xenos, & Larsen, 2003).