Alexandra Balahur: "Hybrid Methods for Sentiment Detection from Social Media"

Ponente: Alexandra Balahur, (Language Techonology Group, Joint Research Centre)

Fecha: martes 26 de febrero de 2013

Hora: 12h00

Lugar de celebración: Sala 1.03, ETSI Informática, UNED

Abstract

In this presentation, I will talk about two of the main lines of research  I am currently pursuing. The first line of research deals with the issue of sentiment classification in multilingual texts, using machine translation systems. I will present different experiments we carried out using a multitude of machine translation systems and different settings for supervised learning, for three different languages. Based on the evaluation results, we propose a set of recommendations for the use of machine translation in the context of multilingual sentiment analysis using supervised learning.

The second line of research is concerned with Social Media mining, especially with linking news about entities to tweets that comment on these news and extracting the opinion expressed on the entities in question. Additionally, we try, in a more general manner, to classify the news in the sense of whether it implies a positive or a negative event for the entity.

To exemplify, I will present the approaches we took and the results we obtained in the CLEF 2012 RepLab profiling (polarity classification) and, more briefly, the monitoring tasks. The tweets in the test set were in English and in Spanish.

Bio

Alexandra Balahur is a member of the Language Technology Group in the Institute for the Protection and Security of the Citizen (IPSC) at the European Commission's Joint Research Centre (JRC). Her main research areas are Sentiment Analysis, Opinion Mining, Emotion Detection and Social Media.

She has recently obtained my PhD in Computer Science, with the thesis entitled "Methods and Resources for Sentiment Analysis in Multilingual Documents of Different Text Types". Her supervisors were Prof. Andres Montoyo and Dr. Ralf Steinberger.

Lugar de celebración

Sala 1.03
ETSI Informática, UNED
c/ Juan del Rosal, 16
Ciudad Universitaria
28040 Madrid

Recursos

 
RocketTheme Joomla Templates