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 find 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.
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 find 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.
M. Atzmueller. Proc. ECML/PKDD 2014: European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, volume 8726 of LNCS, page 485--488. Heidelberg, Germany, Springer Verlag, (2014)
M. Söllner, and J. Leimeister. Workshop on "Revisiting Socio-technical System Design" at the European Conference on Cognitive Ergonomics (ECCE) 2014, Vienna, Austria, (2014)
A. Janson, S. Ernst, K. Lehmann, and J. Leimeister. 4th Workshop on Awareness and Reflection in Technology-Enhanced Learning (ARTEL 2014) to be held in the context of EC-TEL 2014, Graz, Austria, (2014)
S. Niemczyk, R. Kniewel, T. Schulz, and M. Söllner. Socio-technical Design of Ubiquitous Computing Systems, Springer, Berlin (DOI: 10.1007/978-3-319-05044-7_14), (2014)
J. Leimeister, R. Winter, W. Brenner, and R. Jung. Working Paper Services of University of St.Gallen’s Institute of Information Management, No. 1, 1, St. Gallen, Switzerland, (2014)