Matching evaluations and datasets

Heiko Angermann, Naeem Ramzan

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review


The use of heterogeneous taxonomies requires the need of (semi-)automatic information processing and the computation of match scores. Taxonomic heterogeneity occurs in four different categories: terminological heterogeneity (different labels and/or languages are used to describe the concepts), conceptual heterogeneity (contradictory models, including a varying number of hierarchies), syntactical heterogeneity (varying semantic languages used and different syntax), and semiotic heterogeneity (disparate cognitive interpretations and misunderstanding). During the last five years, a large number of matching systems have been proposed, aiming to overcome one or multiple types of taxonomic heterogeneity existing between two taxonomies. The latest best performing matching systems and algorithms now combine multiple matching techniques to ensure the detected alignments, exploit different sources of background knowledge to extract further relations between taxonomy entities, and provide so-called user involvement to correct the resulted correspondences. Based on an analysis of the latest Ontology Alignment Evaluation Initiative (OAEI) results presented for different tracks and test cases between 2011 and 2015, this chapter provides a comprehensive review of state-of-the-art methods and attempts, as well as recent techniques, and discusses open challenges to each of the taxonomic heterogeneity categories.

Original languageEnglish
Title of host publicationTaxonomy Matching Using Background Knowledge
Subtitle of host publicationLinked Data, Semantic Web and Heterogeneous Repositories
Number of pages18
ISBN (Electronic)978-3-319-72209-2
ISBN (Print)978-3-319-72208-5
Publication statusPublished - 8 Jan 2018


Dive into the research topics of 'Matching evaluations and datasets'. Together they form a unique fingerprint.

Cite this