Method for ontology generation from concept maps in shallow domains

Preview:

DESCRIPTION

 

Citation preview

II Jornadas sobre Ontologías y Web Semántica. WebSemántica'2007

Method for ontology generation from concept maps in shallow domains

Alfredo Simón1

Luigi Ceccaroni2 Alejandro Rosete1

1 Technical Institute “José Antonio Echeverría” (Cuba)2 Technical University of Catalonia, Software department (Spain)

asimon@ceis.cujae.edu.cu

Introduction

The development and use of ontologies is increasing today

The design and creation of ontologies, the tools available and the specification languages are still complex (*environment of human collaboration)

These suggests the use of a form of representation that can be used naturally by humans and integrated with ontologies

Integration between an informal, flexible and graphic model of knowledge (*Concept Map) and OWL ontologies

It based on a concept sense-disambiguation procedure and WordNet

It applied to concept maps of shallow domain with labels in the Spanish language

Oriented to obtain OWL-Lite and OWL-DL

Method proposed

Defined by Novak (70’s) within the pedagogical sciences as graphical tools for organize and represent descriptive knowledge

Kind of semantic network but not formal and more flexible

Include concepts and relationship between concepts

Propositions (concept, link-word, concept) are the smallest semantic structure with proper sense

Oriented to be used and interpreted by humans Knowledge is expressed in natural language

Concept Map

Examples (spanish concept maps)

Medical Domain Ontology from the @LIS TechNET Project

It is an ontological and conceptual knowledge representation

It is a sort of semantic network. Therefore similar to Frame Systems

propositions and triples (subject, predicate, object) of RDF have similar structures

structural correspondence:concept with class/subclasslinks and link-word with propertypropositions with restriction and other OWL specifications

Concept maps & OWL ontologies

but it isn’t enough

To increase the formalization levels of the link-words in the concept map

To analyze the concept map as a structured text:Identify the correct sense of the concepts (synset in WordNet)Identify the semantic of the relation between concepts using WordNet (hypernym-hyponym,meronym-holonym)

Five Phases

Formal Transformation

Identification of the Concept Map Domain

Concept Sense Disambiguation for Domain

Concept Sense Disambiguation for Context

Using Spanish WordNet Lexical Database

1st Phase. Concept sense disambiguation

Concept Map Domain

Senses of Arteria concept

1st Phase. Concept sense disambiguation

2nd Phase. Initial coding of OWL classes

3rd Phase. Identification of subclass relations

hypernym relations inhypernym relations in WordNetWordNet

3rd Phase. Identification of subclass relations

4th Phase. Identification of instance relations

5th Phase. Identification of property relations

5th Phase. Identification of property relations

5th Phase. Identification of property relations

5th Phase. Identification of property relations

5th Phase. Identification of property relations

Java application

Input: XML (generated for Macosoft -Concept Maps Editor)

Output: OWL

Validated with Protégé and WonderWeb OWL Ontology Validator

Implementation and Validation

Conclusions

It has been shown that a tight relationship exists between conceptual maps and ontologies

The interpretation of conceptual maps as structured text allows the semantic inference needed for their coding in OWL, without losing flexibility

The defined procedures generate OWL ontologies from conceptual maps

The proposed integration creates the bases for generalization to other domains and for the collaborative development of ontologies

identifications of more OWL specifications inside the concept maps as: TransitiveProperty and SymmetricProperty, property value (hasValue), intersection of class (intersectionOf) and equivalent class

reuse of public ontologies and concept maps repositories

Future Work

Recommended