Social Web Observatory: A Platform and Method for Gathering Knowledge on Entities from Different Textual Sources

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 2000–2008 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 2000 Social Web Observatory: A Platform and Method for Gathering Knowledge on Entities from Different Textual Sources Leonidas Tsekouras∗, Georgios Petasis∗, George Giannakopoulos∗†, Aris Kosmopoulos† ∗Institute of Informatics and Telecommunications, †SciFY PNPC – TEPA Lefkippos National Centre for Scientific Research (N.C.S.R.) “Demokritos”, P.O. BOX 60228, Aghia Paraskevi, GR-153 10, Athens, Greece. {ltsekouras, petasis, ggianna}@iit.demokritos.gr, akosmo@scify.org Abstract Within this work we describe a framework for the collection and summarization of information from the Web in an entity-driven manner. The framework consists of a set of appropriate workflows and the Social Web Observatory platform, which implements those workflows, supporting them through a language analysis pipeline. The pipeline includes text collection/crawling, identification of different entities, clustering of texts into events related to entities, entity-centric sentiment analysis, but also text analytics and visualization functionalities. The latter allow the user to take advantage of the gathered information as actionable knowledge: to understand the dynamics of the public opinion for a given entity over time and across real-world events. We describe the platform and the analysis functionality and evaluate the performance of the system, by allowing human users to score how the system fares in its intended purpose of summarizing entity-centered information from different sources in the Web. Keywords: Knowledge Discovery/Representation, Opinion Mining / Sentiment Analysis, Tools, Systems, Applications 1. Introduction The Social Web Observatory1 is an initiative that aims to help researchers interested in the social sciences and digital humanities study how information spreads in the news and other user-generated content, such as social media posts and comments. The overall system is composed of a back- end and a web application that provides a friendly front-end to the final users. The platform allows users to define their own entities via a simple user interface and can then show a summary of the information that has been gathered about the entity. Entity-driven event detection and summarization is needed in real-life scenarios, such as due diligence, risk assess- ment, fraud detection, etc.; where the entities are usually firms or individuals. In this work we overview Social Web Observatory and we examine, through a human user study, a set of research questions related to its summarization performance: • Are the event clusters created by the system meaning- ful, reflecting a single event? • How well does the system avoid bringing irrelevant articles into the clusters? • Does the system choose representative titles for the identified events? The rest of the paper is structured as follows. In Section 2. we outline some related work and position our work. Then, in Sections 3. and 4. we describe the platform, des- ignate the problem it is meant to face and outline the meth- ods used in the Social Web Observatory analysis pipeline. We continue, in Section 5., by describing the experiments conducted to answer our research questions, which we then discuss in Section 6.. …

Content truncated to 500 words