Campus Event Calendar
Event Entry
What and Who
Multilingual Text Classification using Ontologies
Gerard De Melo
IMPRS-CS
Talk
AG 1, AG 2, AG 3, AG 4, AG 5, SWS
AG Audience
English
Note: We use this to send email in the morning.
Date, Time and Location
Monday, 26 February 2007
08:00
540 Minutes
E1 4
024
Saarbrücken
Abstract
Multilingual text classification is a problem that involves
classifying text documents provided in different languages
thematically, geographically, or according to other criteria.
A novel linguistically motivated strategy called Ontology
Region Mapping will be presented, where formal ontologies and
lexical resources are used to represent text in a way that
enables machine learning algorithms to learn classifications
from pre-classified examples and then automatically classify
documents that might be given in completely different
languages. Ontology Region Mapping associates terms occurring
in a text with concepts represented in formal ontologies and
lexical resources, thereby, however, going beyond a direct
mapping from terms to concepts. In order to fully exploit the
external knowledge manifested by an ontology, entire regions
of relevant concepts are considered, which is achieved by means
of a graph traversal algorithm that explores further related
concepts. Extensive testing has shown that this leads to
significant improvements compared to existing approaches.
classifying text documents provided in different languages
thematically, geographically, or according to other criteria.
A novel linguistically motivated strategy called Ontology
Region Mapping will be presented, where formal ontologies and
lexical resources are used to represent text in a way that
enables machine learning algorithms to learn classifications
from pre-classified examples and then automatically classify
documents that might be given in completely different
languages. Ontology Region Mapping associates terms occurring
in a text with concepts represented in formal ontologies and
lexical resources, thereby, however, going beyond a direct
mapping from terms to concepts. In order to fully exploit the
external knowledge manifested by an ontology, entire regions
of relevant concepts are considered, which is achieved by means
of a graph traversal algorithm that explores further related
concepts. Extensive testing has shown that this leads to
significant improvements compared to existing approaches.
Contact
IMPRS
225
--email hidden
passcode not visible
logged in users only
Jennifer Gerling, 02/21/2007 11:12 -- Created document.