Ontology-based relevance assessment : An evaluation of different semantic similarity measures

• Ontology-based relevance assessment of documents is an importanttask. When viewing this in a business context, approaches arecommonly based on the use of one ontology describing the enterprise. Amore specific problem is then how to assess the relevance of a set of ontologyconcepts with respect to a user profile expressed within the sameontology. Semantic similarity measures have been widely used and describedin literature, but few experiments have been performed to showthe benefits and drawbacks of certain measures. In this paper we describehow a set of measures have been combined, tested, and evaluated. Theevaluation was performed through a rank correlation coefficient comparingthe measured results with a manually constructed “gold standard”.Conclusions are that certain combinations of measures seem less suitablefor solving this type of problem. On the other hand a set of combinationsperform quite well, although the detailed performance of most combinedmeasures seem to depend heavily on the structure of the ontology used.

Subject headings

NATURVETENSKAP  -- Data- och informationsvetenskap (hsv//swe)

NATURAL SCIENCES  -- Computer and Information Sciences (hsv//eng)

Keyword

Ontology

Information technology

Informationsteknik

Publication and Content Type

ref (subject category)

kon (subject category)