finding the most similar concepts in two different ontologies adolfo guzmán-arenas jesús m....
Post on 21-Dec-2015
213 views
TRANSCRIPT
Finding the most Similar Concepts in two Different
Ontologies
Adolfo Guzmán-Arenas
Jesús M. Olivares-Ceja
www.jesusolivares.com [email protected] april 2004
CIC - IPN
Centro de Investigación en Computación
Agenda
Motivation
Ontology definition
Finding the most similar concept in two different ontologies
SIM algorithm
Conclusions and Future work
Motivation
Traditional AI
isolated systems
Current AI
distributed systems
REALITY KnowledgeModel
REALITY
KnowledgeModel
KnowledgeModel
KnowledgeModel
Ontology Definition
An Ontology is an explicit specification of a shared conceptualization (Gruber 1993)
An Ontology consists of a tree of concepts under the subset relation (solid arrow) with other relations (dotted arrow), and words associated to
each concept (in parenthesis)
ornate_plant (ornate plant)
plant_living (plant, vegetal)
animal_living (animal)
farm_animal (...)
savage_animal
(wild animal, beast)
fruit (...)
vegetable (...)
living_thing (creature, live organism, live being)
lion (...)
zebra (...)
chicken (chicken, hen, cock)
eatable_plant(…)
Finding the most similar concept in two different ontologies
Common reference ontology
Ontology BOntology C Ontology A
Top ontology
Passing one concept from one ontology to another
Ontology BOntology A
Finding the most similar concept in two different ontologies
Use the structure within each ontology
CONCEPT (words describing the concept) [PROPERTIES]
words from parent and concept
Ontology BOntology A
agreements
Conclusions and Future Work
Knowledge Exchange among two different previously unknown agents is possible using SIM algorithm
SIM should be tested with real knowledge to be tunned
Different knowledge structures besides of hierarchical (i. e. networks)
Automatic knowledge extraction from different sources
Future Work
Different knowledge structures besides of hierarchical (i. e. networks)
Automatic knowledge extraction from different sources