PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

Improved Semantic Graphs with Word Sense Disambiguation
Delia Rusu, Blaz Fortuna and Dunja Mladenić
In: ISWC 2009, 25-29 Oct 2009, Washington, USA.

Abstract

Semantic graphs can be seen as a way of representing and visualizing textual information in more structured, RDF-like graphs. The reader thus obtains an overview of the content, without having to read through the text. In building a compact semantic graph, an important step is grouping similar concepts under the same label and connecting them to external repositories. This is achieved through disambiguating word senses, in our case by assigning the sense to a concept given its context. The paper presents an unsupervised, knowledge based word sense disambiguating algorithm for linking semantic graph nodes to the WordNet vocabulary. The algorithm is integrated in the semantic graph generation pipeline, improving the semantic graph readability and conciseness. Experimental evaluation of the proposed disambiguation algorithm shows that it gives good results.

EPrint Type:Conference or Workshop Item (Poster)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Computational, Information-Theoretic Learning with Statistics
Learning/Statistics & Optimisation
Information Retrieval & Textual Information Access
ID Code:6464
Deposited By:Jan Rupnik
Deposited On:08 March 2010