Automatic learning of text-to-concept mappings exploiting WordNet-like lexical networks
20th Annual ACM Symposium on Applied Computing Santa Fe, New Mexico, March 13 -17, 2005
KEYWORDS:
Internet,
Semantic Web
ABSTRACT
A great jump towards the advent of the Semantic Web will take place when a critical mass of web resources is available for use in a semantic way. This goal can be reached by the creation of semantic meta-data in the publication workflow, or by the development of systems and applications able to associate semantics to resources (i.e., annotating them) automatically. Those applications should analyze the content of a web page and should be able to associate some ontology classes to it. One particular issue in this context is to define a suitable relationship between each concept of the ontology and some words (or, more in general, strings) which are expected to appear in resources dealing with that concept, playing the role of "triggers" suggesting the relevance of a given text fragment to a concept. We hereby propose an approach that, starting from a set of textual representations created by experts (synsets), is able to auto-matically widen their lexical coverage by computing new, larger synsets, increasing the capability of a semantic application to correctly recognize the ontology classes a document is related to. In such approach, the initial textual representations are integrated and augmented by exploiting lexical networks like WordNet, which contain syntactic information connected through semantic relationships. Some algorithms are proposed to avoid misleading terms and consequently to perform sense disambiguation in WordNet.
| Related files: |
|---|
| sac2005.pdf | Adobe Acrobat portable document |
| sac2005.pdf | Adobe Acrobat portable document [SENSIBLE DATA] |
Notez Bien:
Access to sensible data is granted to domain only. Any use without explicit permission of the CAD group is illegal under the current copyright laws.
Copyright note for papers published by ACM:
Permission to make digital or hard copies of this work for personal or
classroom use is granted without fee provided that copies bear this
notice and the full citation on the first page. Copyrights for
components of this work owned by others than ACM must be
honored. Abstracting with credit is permitted. To copy otherwise, to
republish, to post on servers or to distribute to lists, requires
prior specific permission and/or a fee.
[BCPe05] D. Bonino, F. Corno, F. Pescarmona, "Automatic learning of text-to-concept mappings exploiting WordNet-like lexical networks," 20th Annual ACM Symposium on Applied Computing Santa Fe, New Mexico, March 13 -17, 2005 |
