Automating the generation of semantic annotation tools using a clustering technique

In order to generate semantic annotations for a collection of documents, one needs an annotation schema consisting of a semantic model (a.k.a. ontology) along with lists of linguistic indicators (keywords and patterns) for each concept in the ontology. The focus of this paper is the automatic generation of the linguistic indicators for a given semantic model and a corpus of documents. Our approach needs a small number of user-defined seeds and bootstraps itself by exploiting a novel clustering technique. The baseline for this work is the Cerno project [8] and the clustering algorithm LIMBO [2]. We also present results that compare the output of the clustering algorithm with linguistic indicators created manually for two case studies.