This paper reports on research exploring a threshold for engaging scientists in semantic ontology development. The domain application, nanocrystalline metals, was pursued using a multi-method approach involving algorithm comparison, semantic concept/term evaluation, and term sorting. Algorithms from four open source term extraction applications (RAKE, Tagger, Kea, and Maui) were applied to a test corpus of preprint abstracts from the arXiv repository. Materials scientists identified 92 terms for ontology inclusion from a combined set of 228 unique terms, and the term sorting activity resulted in 9 top nodes. The combined methods were successful in engaging domain scientists in ontology design, and give a threshold capacity measure (threshold acceptability) to aid future work. This paper presents the research background and motivation, reviews the methods and procedures, and summarizes the initial results. A discussion explores term sorting approaches and mechanisms for determining thresholds for engaging scientist in semantically-driven ontology design and the concept of ontological empowerment.
CITATION STYLE
Greenberg, J., Zhang, Y., Ogletree, A., Tucker, G. J., & Foley, D. (2015). Threshold determination and engaging materials scientists in ontology design. In Communications in Computer and Information Science (Vol. 544, pp. 39–50). Springer Verlag. https://doi.org/10.1007/978-3-319-24129-6_4
Mendeley helps you to discover research relevant for your work.