Automatic Hierarchical Clustering of Web Pages
Ricardo Campos, Gaël Dias · 2005
In this paper, we propose a system that clusters web pages and presents them as a hierarchical structure instead of a classical search ordered list retrieved from any search engine. The organization of the results based on this concept makes easier the user’s search navigation between the results of the search engine. In particular, we use web content mining techniques to represent texts, based on their most relevant terms, which can be simple words or phrases. A soft clustering algorithm is then applied to group documents into clusters hierarchically linked. Finally, each cluster is labelled with its most relevant term based on a simple heuristics.