SEMANTAGS: SEMANTIC TAGGING OF DOCUMENTS BASED ON CROWDSOURCING

Enriching objects (e.g., Internet resources) with semantic labels (or, tags) characterizing the content of these objects allows to significantly improve the search results quality and, in general, contributes to more efficient processing of information. However, fully automatic semantic tagging does not always give an adequate result due to existing constraints in the algorithmic processing of natural language, while manual tagging is rather laborious, especially if the number of objects is large. The latter can be partly alleviated in some cases by using the crowdsourcing technology. The article proposes SemanTags system (web-service), which allows to leverage crowdsourcing to tag arbitrary objects (scientific articles, Internet pages, etc.) with classes of some problem-oriented ontology written in OWL 2. The paper describes the architecture of the system, as well as the new quality assurance mechanism, adapted for social tagging of objects by classes of an ontology.

Authors: I. A. Shcherbakov , A. V. Ponomarev

Direction: Informatics, Computer Technologies And Control

Keywords: Crowdsourcing, crowd computing, ontologies, OWL 2, taxonomy, semantic tagging, semantic search


View full article