Adaptation and Evaluation of a Semantic Similarity Measure for DBPedia: A First Experiment
Valentin Grouès, Yannick Naudet, Odej Kao · 2012
Semantic similarity measures aim at computing a value that represents to what extent two semantic resources are similar. This paper proposes an extension of a semantic similarity measure for large linked data sources such as DBPedia. In order to evaluate and compare similarity measures, an experiment is described to collect human evaluations of the similarity between pairs of movies. A correlation of 0.69 is observed between the human ratings and the results of the semantic similarity measure.