Unsupervised Word Polarity Tagging by Exploiting Continuous Word Representations

Autores: Aitor García Pablos Montserrat Cuadros Oller German Rigau Claramunt

Fecha: 16.09.2015

Revista de Procesamiento de Lenguaje Natural


PDF

Abstract

Sentiment analysis is the area of Natural Language Processing that aims to determine the polarity (positive, negative, neutral) contained in an opinionated text.  A usual resource employed in many of these approaches are the so-called polarity lexicons.  A polarity lexicon acts as a dictionary that assigns a sentiment polarity value  to  words.   In  this  work  we  explore the possibility  of  automatically  generating domain adapted polarity lexicons employing continuous word representations, in particular the popular tool Word2Vec.  First we show a qualitative evaluation of a small set of words, and then we show our results in the SemEval-2015 task 12 using the presented method.

BIB_text

@Article {
title = {Unsupervised Word Polarity Tagging by Exploiting Continuous Word Representations},
journal = {Revista de Procesamiento de Lenguaje Natural},
pages = {127-134},
volume = {55},
keywds = {

word embeddings, polaridad de palabras, análisis de sentimiento


}
abstract = {

Sentiment analysis is the area of Natural Language Processing that aims to determine the polarity (positive, negative, neutral) contained in an opinionated text.  A usual resource employed in many of these approaches are the so-called polarity lexicons.  A polarity lexicon acts as a dictionary that assigns a sentiment polarity value  to  words.   In  this  work  we  explore the possibility  of  automatically  generating domain adapted polarity lexicons employing continuous word representations, in particular the popular tool Word2Vec.  First we show a qualitative evaluation of a small set of words, and then we show our results in the SemEval-2015 task 12 using the presented method.


}
date = {2015-09-16},
year = {2015},
}
Vicomtech

Parque Científico y Tecnológico de Gipuzkoa,
Paseo Mikeletegi 57,
20009 Donostia / San Sebastián (España)

+(34) 943 309 230

close overlay