Abstract
Digitization of cultural heritage collections, both within and outside museums has led to the establishment of a new field of theory and practice of the digital curation of cultural information. Cultural heritage ontology construction a practice in museum documentation, faces the challenge of dealing with the abundance of information while struggling to maintain the authenticity and preservation of the cultural knowledge. Similar to other knowledge domains, the existence of various ontology of cultural heritage knowledge and their differences has caused predicament in accessing or retrieving information. To overcome this predicament, the standard ontology, in this case the CIDOC CRM is used to consolidate between two or more local ontology through the consolidation with a standard (global) ontology through a process known as ontology mapping. The effort needed to map two ontologies with some content similarity but different structures while minimizing knowledge loss is challenging. Knowledge loss is a situation where the concepts of local ontology are excluded from the result of mapping to the global ontology. Automated mapping through the use various existing algorithms offers an efficient solution but suffers in knowledge loss. A compromised method is needed to balance between the construction efficiency and the knowledge loss. This research seeks to establish a framework for constructing cultural heritage ontology through semi-automated ontology mapping. Using the traditional Malays textile (TMT) as the domain knowledge for this work and scoping on the Malaysian batik, this research is carried out in three phases. In the first phase, a manual mapping process between an existing TMT Knowledge Model (local) and CIDOC CRM (global) was carried to produce a Malaysian batik heritage ontology (MBHO). This is also achieved due to active participation of batik and ontology experts in verifying the MBHO. In the second phase, several automated mapping tools were tested. The finding shows that the mapping tools able to produce a Malaysian Batik ontology close to MBHO with knowledge loss as the resulting mapping were mostly incorrect. In the third phase, the refinement of the processes undergone in phases one and two were made to deduce the actual steps carried out and the rules that govern the actions of ontology construction through ontology mapping process. Comparison between both processes showed that the semi-automated mapping can improve the efficiency of the construction while reducing knowledge loss. The work in this phase led to the construction of a process framework of ontology construction through semi-automation. This framework is then verified by an expert in framework development and two experts in ontology constructions. A revised framework was produced and is name as the Framework for Ontology Construction through Semi-Automated Mapping (MapOn). This new framework contributes towards the knowledge and practice of digital curation in general and ontology construction in specific. In addition the MBHO is a new ontology for the Malaysian cultural heritage. This research also produces empirical evidence of the phenomena of knowledge loss through automated mapping which supports the inclination towards semi-automated mapping.
Metadata
Item Type: | Thesis (PhD) |
---|---|
Creators: | Creators Email / ID Num. Md Nasir, Syerina Azlin UNSPECIFIED |
Divisions: | Universiti Teknologi MARA, Shah Alam > Faculty of Computer and Mathematical Sciences |
Keywords: | Process framework; Ontology construction; Ontology mapping |
Date: | 2016 |
URI: | https://ir.uitm.edu.my/id/eprint/16795 |
Download
TP_SYERINA AZLIN MD NASIR CS 16_5.pdf
Download (7MB) | Preview