Method for Ontology-based Case Based Reasoning (tool OntoCase)
One sentence description of the method.
Institution: | Institute of Informatics |
Technologies used: | Java, JENA |
Inputs: | Ontology-based Query Case, Profile |
Outputs: | Expected Case Result |
Video: | demonstration video |
Distribution packages: | zip |
Addressed Problems
Description
OntoCase uses Case-Based Reasoning (CBR) approach to solve new problems based
on the solutions of similar past problems. Case is a solution to given problem,
where problem is a set of “feature – value” pairs and solution is a “result”:
{N, V, R } e C
Case Base is a set of all previous (historically) observed cases or set of
generated a reused cases.
Ontology-based CBR
In case of Ontology-based Case-Based Reasoning resources are described in the ontology and properties of instances represent features. Features can be not only continuous or discrete but also Concepts (instance-of owl:Class) or Instances (instance-of: owl:Thing) in the ontology.

Overal OntoCase Approach
Case - is a basic concept in OCBR. Case concept has Context (which provides the problem description) and Result (concept or instance describing the result of the problem/case). Context is abstract and needs to be extended.

Ontology-based Case Representation
Profile - we need to specify what to estimate (which Result concepts should be estimated) and which properties (WS invocation parameters and Resource properties) should be the part of the feature vector. For such purposes we are proposing to use a Profile concept.

Profile
Feature - enables us to specify which property of a Resource to use as a feature during the retrieval process. WeightedFeature - enables to add weight to a feature.


Feature and Weighted Feature
Tool OntoSim was used to introduce structural similarity of concepts in ontology to refine the retrieval process of IBL. We have reached better results using Object Feature as when using just Numerical Features.
References
- Balogh Z., Budinska I., Hluchy L.: OntoCase – prípadové usudzovanie založené na ontologiách In: Proceedings 2nd Workshop on Intelligent and Knowledge Oriented Technologies - WIKT 2007 proceedings. - Košice : Centre for Information Technologies, Faculty of Electrical Engineering an Informatics, Technical University in Košice, 2008. ISBN 978-80-89284-10-8, p. 81-85.