Social bookmarking systems and their emergent information structures, known as folksonomies, are increasingly important data sources for Semantic Web applications. A key question for harvesting semantics from these systems is how to extend and adapt traditional notions of similarity to folksonomies, and which measures are best suited for applications such as navigation support, semantic search, and ontology learning. Here we build an evaluation framework to compare various general folksonomy-based similarity measures derived from established information-theoretic, statistical, and practical measures. Our framework deals generally and symmetrically with users, tags, and resources. For evaluation purposes we focus on similarity among tags and resources, considering different ways to aggregate annotations across users. After comparing how tag similarity measures predict user-created tag relations, we provide an external grounding by user-validated semantic proxies based on WordNet and the Open Directory. We also investigate the issue of scalability. We ?nd that mutual information with distributional micro-aggregation across users yields the highest accuracy, but is not scalable; per-user projection with collaborative aggregation provides the best scalable approach via incremental computations. The results are consistent across resource and tag similarity.

@inproceedings{markines2009evaluating, author = {Markines, Benjamin and Cattuto, Ciro and Menczer, Filippo and Benz, Dominik and Hotho, Andreas and Stumme, Gerd}, title = {Evaluating Similarity Measures for Emergent Semantics of Social Tagging}, booktitle = {18th International World Wide Web Conference}, year = {2009}, pages = {641--641}, url = {http://www.kde.cs.uni-kassel.de/pub/pdf/markines2009evaluating.pdf}, keywords = {2009, itegpub, myown, social_similarity, tagorapub, www2009, ol_web2.0, semantic_relatedness, methods_concepts}, abstract = {Social bookmarking systems and their emergent information structures, known as folksonomies, are increasingly important data sources for Semantic Web applications. A key question for harvesting semantics from these systems is how to extend and adapt traditional notions of similarity to folksonomies, and which measures are best suited for applications such as navigation support, semantic search, and ontology learning. Here we build an evaluation framework to compare various general folksonomy-based similarity measures derived from established information-theoretic, statistical, and practical measures. Our framework deals generally and symmetrically with users, tags, and resources. For evaluation purposes we focus on similarity among tags and resources, considering different ways to aggregate annotations across users. After comparing how tag similarity measures predict user-created tag relations, we provide an external grounding by user-validated semantic proxies based on WordNet and the Open Directory. We also investigate the issue of scalability. We ?nd that mutual information with distributional micro-aggregation across users yields the highest accuracy, but is not scalable; per-user projection with collaborative aggregation provides the best scalable approach via incremental computations. The results are consistent across resource and tag similarity.} }

%0 = inproceedings %A = Markines, Benjamin and Cattuto, Ciro and Menczer, Filippo and Benz, Dominik and Hotho, Andreas and Stumme, Gerd %B = 18th International World Wide Web Conference %D = 2009 %T = Evaluating Similarity Measures for Emergent Semantics of Social Tagging %U = http://www.kde.cs.uni-kassel.de/pub/pdf/markines2009evaluating.pdf