In this talk we show that LOD annotations are a powerful tool for semantic enrichment of Social Media microposts, allowing for reasoning with information that is not transmitted directly through the Social Media channels, but available in rich knowledge bases. Given the very short text of tweets, such enrichment provides with the necessary context, which is crucial for understanding opinions, trends, veracity in Social Media. LOD enrichment allows for the computer algorithms to ‘understand’ tweets in a way that a human would, by referring to common knowledge external to the micropost.
We choose a showcase that is highly relevant for the international political scene at the moment of the deliverable, namely a post-Brexit analysis. Post-Brexit discussions on Twitter provide with insights on the mixed feelings, attitude, propaganda, interests that follow the referendum and precede the political actions that need to be taken. We show that one can automatically mine the general opinion of the main UK administrative regions. We also identified the main actors of the political scene, with side comments on their age - an aspect that has been so many times brought to the public attention and even used for manipulating the opinion of the voters. Reasoning about age or birth year is only possible via LOD annotations.
Vladimir Alexiev has a PhD in computing science from University of Alberta, MS in computer science from Technical University of Sofia, PMP certification, Project Risk and Quality Management diploma. His semantic web experience includes ontology engineering, metadata standards, vocabularies and thesauri, RDF, RDFS, OWL2, SHACL, SKOS, SPARQL, LOD, mapping, R2RML, ETL. At Ontotext he leads a wide range of projects related to cultural heritage, archives, libraries, CIDOC CRM, Europeana, LOD management, thesauri, etc. He served on 5 Europeana task forces (including 2 on semantic enrichment), the Members Council and the Data Quality Committee. He is on the DBpedia Ontology and Data Quality committee, seeking to improve the quality of DBpedia ontology, mappings and data. He has contributed to the ontology definition of the ISO 25964 standard for thesauri management; on appropriate composition of different "Broader Than" relations (BTG, BTP, BTI); on publishing important CH thesauri as LOD. He has particular interest in the faithful representation of humanities data (museum, library, archaeology, medieval studies, etc) as RDF, CIDOC CRM and its extensions, practical applications with such data.
Lead, Data and Ontology Management