China
Africa
Explore chapters and articles related to this topic
1
Published in Paul Compton, Byeong Ho Kang, Ripple-Down Rules, 2021
The influence of Plato seems clear in modern work on ontologies. There is no doubt that it is useful to have a set of terms to be used consistently – and even better to have a formal structure for the concepts and relationships the terms express. But modern work on ontologies also includes the development of so-called upper ontologies which cover ideas like object, quality, attribute, property, quantity, processes, events, and the ultimate goal is a shared universal ontology enabling reasoning across the Semantic Web. Despite the importance of developing shared standard ontologies, it is not at all clear whether they will be able to be used by humans. The Unified Medical Language System (UMLS) (not actually an ontology) has over one million concepts, five million synonyms and hundreds of terminologies. Rosenbloom et al. have pointed out that the most important issue with the UMLS is not its completeness, but how end-users can be enabled to use it (Rosenbloom et al. 2006). A recent study of some biomedical ontologies in use found significant errors (Boeker et al. 2011). Along with errors in ontologies, a further issue is their multiplicity so that ontology merging is a major research issue, but the challenge is not simply that different terminologies have emerged but that people disagree even on ontology alignment (Tordai et al. 2011). On the other hand, Tordai et al. comment: “Humans rarely have problems with disambiguating the meaning of words in a discourse context”. In an earlier simpler example Shaw and Woodward carried out a study comparing different domain experts, as well as repeat studies with the same experts as a control (Shaw and Woodward 1988). They found experts construct knowledge with different terminologies and disagree with each other’s terminologies but that this is not a problem as they are well used to both working together and disagreeing with each other!
Ontology Modeling
Published in Archana Patel, Narayan C. Debnath, Bharat Bhushan, Semantic Web Technologies, 2023
Ontolingua, often considered an ontology language, is an ontology authoring and management tool large variety of user preferences for controlling the behavior of the user to facilitate the development and sharing of ontologies. It is a web-based ontology editor that aids with basic development tasks, facilitate collaboration, and ease of use. It provides a mechanism for writing ontologies in a traditional format so that they can be easily translated into a variety of representation and reasoning systems. Ontolingua system consists of a suite of tools for authoring ontologies from existing fragments of ontologies. The tool supports ontology inclusion, circular dependencies, and polymorphic refinement. It enables the creation and separation between the presentation and representation of ontology. One mechanism employed by Ontolingua in eliminating conflict arising from ontology merging is by making the symbol vocabulary of every ontology disjoint from the symbol vocabulary of all other ontologies. As earlier reported in some ontology tools, Ontolingua also provides users with features for collaboration. The collaborative work is facilitated by user and group access control through multi-user sessions. The collaborative nature of Ontolingua is made obvious by its use of the world-wide web to enable wide access and through the provisioning for users with the ability to publish, browse, create, and edit ontologies stored on an ontology server. Unlike other ontology editors, Ontolingua Server currently does not provide many inferential capabilities nor does it have some plugins to aid ontology reasoning. However, it does provide some support for using ontologies. One way to use ontologies developed with the Ontolingua Server is to translate the ontology into the representation language of another system such as CLIPS, LOOM, or Prolog. Ontolingua editor enables users to be able to browse, edit, publish, and collaborate on ontologies. The browsing environment is being able to quickly jump from one term in the ontology to another through the aid of hyperlinks. Each term or concept represented in the interface is displayed in an object-oriented or frame-based form. Also, Ontolingua Server provides features to assist with ontology maintenance. Another feature enabled in Ontolingua is support for splitting a large ontology into several smaller ontologies that may include each other. Ontology sharing is another feature in Ontolingua with the primary mechanism for supporting ontology reuse.
Ontological knowledge integration and sharing for collaborative product development
Published in International Journal of Computer Integrated Manufacturing, 2018
Xiuzhen Li, Zhenyong Wu, Mark Goh, Siqi Qiu
When the concepts from two ontologies are similar, the concepts should be merged (Noy and Musen 2003; Kotis, Vouros, and Stergiou 2006). In the ontology merging process, similar attributes and concepts from the local ontology are integrated into the global ontology, as shown in Figure 2. First, the similarities of attributes between the local and global ontologies are calculated by attribute mapping. If the attribute similarity is less than 0.5, then the attribute from the local ontology is copied into the global ontology, and a new attribute node is added to store the copied attribute in the new global ontology. Otherwise, the attribute is merged into the global ontology. Second, the similarities of the concepts between the local ontology and the global ontology are measured with Equation (3). Similarly, if the concept similarity is less than 0.8, the attributes and concept from the local ontology are copied into the global ontology, and the new attribute and concept nodes are stored in the new global ontology. Otherwise, the similar attributes and concept are merged into the global ontology. The global ontology is then updated.
CLOE: a cross-lingual ontology enrichment using multi-agent architecture
Published in Enterprise Information Systems, 2019
Mohamed Ali, Said Fathalla, Shimaa Ibrahim, Mohamed Kholief, Yasser F. Hassan
A recent review of the literature on ontology learning found that most studies tended to focus on learning ontologies from monolingual data sources for domain-specific ontologies, i.e domain-dependent (Table 1). Learning ontologies from multilingual data sources using a domain-independent approach has become an intriguing area in the field of ontology learning. Cross-lingual ontology mapping approaches are used for the automatic construction and enrichment of multilingual or large lexical ontologies (De Melo and Weikum 2009; Helou and Palmonari 2015). Since more than a million datasets have now been published online as linked open data (LOD), from 43 countries in 24 different natural languages (IOGDS 2017), cross-lingual ontology mapping has become a challenging area in the field of ontology learning. Park, Cho, and Rho (2010) have divided the ontology learning tools into three categories: (1) Ontology editing tools, which provide the help to ontology engineers for acquiring, organizing, and visualizing domain knowledge; (2) Ontology merging tools, which create one coherent ontology from two or more existing ontologies; and (3) Ontology enrichment tools, which extract concepts and/or relations from unstructured data by using NLP and/or machine learning techniques. Yang and Callan (2008) presented a technique to extract concepts from a corpus of public comments. Candidate concepts were identified by n-gram mining. WordNet (Fellbaum 1998) and surface text pattern matching (Ravichandran and Hovy 2002) are used to identify relationships among the concepts and to guide the organization of concepts into intended hierarchical relationships.