Khan, Zubeida C and Keet, C. Maria (2017) Automatically changing modules in modular ontology development and management, Proceedings of Annual Conference of the South African Institute of Computer Scientists and Information Technologists (SAICSIT'17), 26-28 Sept 2017, Thaba Nchu, South Africa, 19:1-19:10, ACM.
PDF
ModuleSwapSAICSIT17.pdf Download (805kB) |
Abstract
Modularity has been proposed as a solution to deal with large ontologies. This requires, various module management tasks, such as swapping an outdated module for a new one or a computationally costly one for a leaner fragment. No mechanism exists to exchange an arbitrary module automatically. To solve this manual task, we modify the SUGOI algorithm into SUGOI-Gen; with SUGOI-Gen, one can swap any module within a modular system, implemented it, and wrapped a GUI around it. We carried out an experimental evaluation with six ontologies covering three different use-cases to determine whether arbitrary interchangeability is practically doable, and to what extent such changes affect the quality of the module and automated reasoning over it. The results are positive, with the success rate varying between 22-100% depending on the number of mappings between the source and target module. The evaluation also revealed that the interchangeability does indeed have an impact on a module’s metrics. Regarding reasoning, when comparing an original ontology to one where a module has been swapped, the processing time is greatly improved for all except one of the swapped modules in the set.
Item Type: | Conference paper |
---|---|
Uncontrolled Keywords: | ontology engineering, modular ontologies, OWL |
Subjects: | Computing methodologies > Artificial intelligence |
Alternate Locations: | http://www.meteck.org/files/ModuleSwapSAICSIT17.pdf, https://dl.acm.org/citation.cfm?id=3129443 |
Date Deposited: | 18 Jan 2018 |
Last Modified: | 10 Oct 2019 15:31 |
URI: | http://pubs.cs.uct.ac.za/id/eprint/1235 |
Actions (login required)
View Item |