Semantic Links Using SKOS Predicates

Presentation in the 21st International Conference on Knowledge-Based and Intelligent Information & Engineering Systems (KES2017)

Transcript

1. KES 2017 Semantic Links Using SKOS Predicates 0/16 Semantic Links Using SKOS Predicates Ricardo Ávila 8th September 2017 David Araújo Gabriel Lopes Vania Vidal José Macedo Federal University of Ceará, Brazil
2. KES 2017 Semantic Links Using SKOS Predicates 0/16 Outline Contextualization and objectives of this work Problem LDM Generation Process LDM Execution Environment Proposed Methodology Results and Contributions Conclusion and Future Work
3. KES 2017 Semantic Links Using SKOS Predicates 0/16 • Linked Data Mashups (LDM) are interactive Web services that combine the content of different data sources into a new service • Semantic Web provides technologies that enable the publishing, retrieval, and integration of distributed data in the Data Web • Developing link between heterogeneous sources using conventional programming languages is a complex task • Linked Data enables access to data that is semantically related • This work present a framework based on ontologies in order to facilitate semantic integration between different glossaries Contextualization
4. KES 2017 Semantic Links Using SKOS Predicates 0/16 • Would it be possible not to use domain experts to solve simple problems? • Linked Data Mashup would help non-specialists to solve simple problems? Problem
5. KES 2017 Semantic Links Using SKOS Predicates 0/16 LDM Generation Process • The process of creating an LDM initially focuses on the modeling of a domain ontology and semantic integration. (Tran et al. 2014) Figure 1. Domain Ontology (DO) of the Glossary Mashup
6. KES 2017 Semantic Links Using SKOS Predicates 0/16 LDM Execution Environment Figure 2. Source Ontologies (SOs) of the Glossary Mashup
7. KES 2017 Semantic Links Using SKOS Predicates 0/16 Proposed Methodology (1/4) Table 1. Glossaries of Terms in the Oil Domain
8. KES 2017 Semantic Links Using SKOS Predicates 0/16 Proposed Methodology (2/4) Figure 3. Instance Example
9. KES 2017 Semantic Links Using SKOS Predicates 0/16 Proposed Methodology (3/4) • For automatic generation of SKOS predicates, we developed an algorithm to validate and evaluate the performance of the proposed method, based on the predicate skos:prefLabel Figure 4. Developed Algorithm
10. KES 2017 Semantic Links Using SKOS Predicates 0/16 These types of relationships are represented according to the following definitions: •ET (Exact Terms), defined between two terms ti and tj, provided that they are considered the same word. ET is symmetric, that is, ti ET tj ! tj ET ti. •BT (Broader Terms), defined between two terms ti e tj, provided ti has a more general meaning than tj. BT is not symmetrical. •NT (Narrower Terms) is opposite of BT: ti NT tj ! tj BT ti. •RT (Related Terms), defined between two terms ti e tj, which are generally used together in the same context. RT is symmetric: ti RT tj ! tj RT ti. Proposed Methodology (4/4)
11. KES 2017 Semantic Links Using SKOS Predicates 0/16 Results Table 2. Number of SKOS predicates generated by the algorithm Table 3. Examples of SKOS predicates generated by the algorithm
12. KES 2017 Semantic Links Using SKOS Predicates 0/16 • In this research work, some important considerations could be made: • elaboration of the conceptual model of the application in a Domain Ontology (DO) • methodology to mapping correspondences between DO and Source Ontology (SO) • An effective algorithm for automatic generation of skos:exactMatch predicates Conclusion
13. KES 2017 Semantic Links Using SKOS Predicates 0/16 • Vocabulary that can be adapted to other domains • In complexity level, a simple solution for generation of semantic links (LDM) • Any glossary, in any language, can be used as a data source Contributions
14. KES 2017 Semantic Links Using SKOS Predicates 0/16 • The performance of the proposed methodology was less effective in the other predicates skos:broader, skos:narrower and skos:related. To mitigate these problems, we propose: • use the synonym of the terms, machine learning and / or mapping of the classes and subclasses of the domain • Improve results of terms, especially for generation of semantic links in cases of synonyms, hyperonyms and hyponyms • Testing with larger samples and in other contexts • Create rules to other predicates: skos:broaderTransitive, skos:broadMatch, skos:narrowerTransitive, skos:narrowMatch and skos:closeMatch Future Work
15. KES 2017 Semantic Links Using SKOS Predicates 0/16 1. Avila, R. and Soares, J. M. (2012). Concepção de Ferramenta de Apoio à Correção de Questões Dissertativas com Base na Adaptacão de Algoritmos de Comparação e Busca Textual Combinados com Técnicas de Pré-Processamento de Textos. In RENOTE Revista Novas Tecnologias na Educação, volume 10. 2. Bizer, C., Cyganiak, R., and Gauß, T. (2007). The RDF Book Mashup: From WEB APIS to a Web of Data. In ESWC’07 Workshop on Scripting for the Semantic Web, volume 1. 3. d’Aquin, M., Kronberger, G., and Suárez-Figueroa, M. C. (2012). Combining Data Mining and Ontology Engineering to Enrich Ontologies and Linked Data. In KNOW@LOD, volume 868 of CEUR Workshop Proceedings, pages 19–24. CEUR-WS.org. 4. Ge, J. and Chen, Z. (2010). Constructing Ontology-Based Petroleum Exploration Database for Knowledge Discovery. Trans Tech Publications, Switzerland, 20-23:975–980. 5. Heath, T. and Bizer, C. (2011). Linked Data: Evolving the Web into a Global Data Space. Morgan & Claypool, 1st edition. 6. Kazi, A. and Kurian, D. (2014). An Ontology-Based Approach to Data Mining. In International Journal of Engineering Development and Research (IJEDR), volume 2. 7. Michael, J., Mejino Jr., J., and Rosse, C. (2001). The Role of Definitions in Biomedical Concept Representation. Proceedings Annual Symposium. AMIA, pages 463–467. Cited By 13. 8. Miles, A., Matthews, B., Wilson, M., and Brickley, D. (2005). SKOS Core: Simple Knowledge Organization for the WEB. In Proceedings of the 2005 International Conference on Dublin Core and Metadata Applications: Vocabularies in Practice, DCMI ’05, pages 1:1–1:9. 9. Miranda, S., Orciuoli, F., and Sampson, D. G. (2016). A SKOS-Based Framework for Subject Ontologies to Improve Learning Experiences. Comput. Hum. Behav., 61(C):609–621. 10. Rahm, A. T. D. A. E., Thor, D., and Aumueller, E. (2007). Data Integration Support for Mashups. In Workshops at the Twenty-Second AAAI Conference on Artificial Intelligence. 11. Tran, T. N., Truong, D. K., Hoang, H. H., and Le, T. M. (2014). Linked Data Mashups: A Review on Technologies, Applications and Challenges, pages 253–262. Springer International Publishing, Cham. References
16. KES 2017 Semantic Links Using SKOS Predicates 0/16 Thank you!