|
One of the main topics of my research is acquisition of lexical
knowledge for refining ontologies, and it estudies several possible
unsupervised methods towards that aim. Several experiments have been
performed to extend WordNet with new concepts automatically extracted
from mythological texts, such as Homer's The Iliad and
Tolkien's The Lord of the Rings. The procedures have been
applied to hypermedia education generation by adapting WordNet to the
domains of on-line information sites about Darwin'sThe Voyages of
the Beagle Osler's treaty about the history of medicine, and
Hegel's Lectures on the history of philosophy.
The product of this work is the Welkin system, which is able
to automatically generate adaptive on-line information sites. The
final aim is to show each user the portions of the original texts in
which s/he is more interested.
Welkin combines techniques from Natural Language Processing with
Adaptive Hypermedia so as to adapt linear texts into adaptive web
sites. The off-line processing has been designed as a pipeline of
components that perform various linguistic analyses, identify
terminology, and automatically classify the terms inside the
conceptual semantic network WordNet. During the on-line step, when
users connect the on-line site, the system keeps personalised interest
profiles and filters and summarises the contents of the site
accordingly. Finally, it is possible to read additional information
retrieved automatically from Internet about topics related to the
on-line sites.
The system is accessible for those connecting inside the
Universidad Autónoma de Madrid, here.
|
To download the thesis, click here
|
|