Multilingual Lexicalisation and Population of Event Ontologies: A Case Study for Social Media

In this book chapter we will describe a semi-automatic method for ontology-driven lexical acquisition and ontology population. The method is based on a multilingual weakly-supervised approach for learning of semantic classes and event patterns from unannotated text corpus. To illustrate the feasibility of our approach, we apply it on a corpus of Twitter messages for two languages and we populate a micro-ontology of events in these languages. Our proposal is about multi- and cross-lingual ontology-based information extraction and ontology population. We propose it for the section Methods. The proposed method can be viewed as a step towards the development of Social Semantic Web [3]. In the Social Semantic Web, user generated content is labeled with ontology-based semantics, linking in this way user communities from different social networks. Our lexical learning and ontology population approach can be used to automatically annotate or propose semantic annotation to status messages, written by social media users. This potential application is similar to the approach described in [11]. We already experimented with learning and populating a prototype of a micro-ontology from Twitter. We evaluated the extracted information and we found the preliminary results encouraging.

TANEV Hristo;  ZAVARELLA Vanni; 

2017-09-13

Springer Verlag

JRC77535

978-3-662-43584-7,   

https://link.springer.com/chapter/10.1007%2F978-3-662-43585-4_16,    https://publications.jrc.ec.europa.eu/repository/handle/JRC77535,   

10.1007/978-3-662-43585-4_16,