Linked Open Vocabulary Recommendation Based on Ranking and Linked Open Data

摘要

The vocabulary space of the Semantic Web includes more than 500 vocabularies according to the Linked Open Vocabularies (LOV) initiative that maintains the directory list and provides search functionality on top of the curated data. Domain experts and researchers have populated it to facilitate the interpretation and exchange of information in the Web of Data. The abundance of vocabularies and terms available in the LOV space, on one hand aims to cover the major knowledge management needs, but on the other hand it could be cumbersome for a non-expert or even a vocabulary expert to find the correct way through the collection. To address this problem, we present an approach that helps to identify the most appropriate set of LOV vocabulary terms for a given Web content context by leveraging the existing dynamics within the LOV graph and the usage patterns in the LOD cloud. The paper describes the framework architecture that enables the discovery of vocabularies; it focuses on the corresponding metrics and algorithm, and discusses the outcomes of the applied experiments.