Charla: “Supervised models for Twitter opinion lexicon expansion”


Compartir
Charlista: 
Felipe Bravo-Marquez, University of Waikato
Fecha: 
20 Agosto, 2015 - 11:00
Sala: 
Auditorio DCC, Piso 4
Organización: 
Benjamin Bustos

Abstract: 

Many sentiment analysis applications rely on opinion lexicons, which are  linguistic resources that associate words to sentiment values.  Twitter-specific sentiment applications must deal with a language that  includes many informal expressions that are not observed in traditional  media, e.g., acronyms, hashtags, misspelled words, and abbreviations. The  diversity and sparseness of this language make the manual creation of a  Twitter-oriented opinion lexicon a time-consuming task. In this talk, we  will present two different supervised models for automatically discovering  Twitter opinion words from a corpus of tweets. Both works were recently  published at IJCAI and SIGIR conferences respectively. 

 

Felipe Bravo-Marquez is currently doing his PhD at the machine learning  group in the University of Waikato, New Zealand. He received two engineering  degrees in the fields of computer science and industrial engineering, and a  master’s degree in computer science, all from the University of Chile. He  worked for three years as a research engineer at Yahoo! Labs Latin America.  His main areas of interest are: data mining, natural language processing,  information retrieval, and sentiment analysis.

You can find his full list of publications at:  http://www.cs.waikato.ac.nz/~fjb11/