Skip to main content

Resultados relacionados

Resultados totales (Incluyendo duplicados): 9
Encontrada(s) 1 página(s)

RUC. Repositorio da Universidade da Coruña
Publication Contribución de congreso (ConferenceOutput)

EN-ES-CS: An English-Spanish Code-Switching Twitter Corpus for Multilingual Sentiment Analysis

  • Vilares, David
  • Alonso, Miguel A.
  • Gómez-Rodríguez, Carlos
[Abstract]: Code-switching texts are those that contain terms in two or more different languages, and they appear increasingly often in social media. The aim of this paper is to provide a resource to the research community to evaluate the performance of sentiment classification techniques on this complex multilingual environment, proposing an English-Spanish corpus of tweets with code-switching (EN-ES-CS CORPUS). The tweets are labeled according to two well-known criteria used for this purpose: SentiStrength and a trinary scale (positive, neutral and negative categories). Preliminary work on the resource is already done, providing a set of baselines for the research community., This research is supported by the Ministerio de Economía y Competitividad (FFI2014-51978-C2) and Xunta de Galicia (R2014/034). David Vilares is funded by the Ministerio de Educación, Cultura y Deporte (FPU13/01180). Carlos Gómez-Rodríguez is funded by an Oportunius program grant., Xunta de Galicia; R2014/034
RUC. Repositorio da Universidade da Coruña
Publication Contribución de congreso (ConferenceOutput)

LyS at TASS 2014: A Prototype for Extracting and Analysing Aspects from Spanish tweets, LyS en TASS 2014: Un prototipo para la extracción y análisis de aspectos en tuits.

  • Vilares, David
  • Doval, Yerai
  • Alonso, Miguel A.
  • Gómez-Rodríguez, Carlos
[Abstract]: This paper describes our participation at the third edition of the work- shop on Sentiment Analysis focused on Spanish tweets, tass 2014. This year’s eval- uation campaign includes four challenges: (1) global sentiment analysis, (2) topic classification, (3) aspect-extraction and (4) aspect-based sentiment analysis. Tasks 1 and 2 are addressed from a machine learning approach, using several linguistic resources and other information extracted from the training corpus to feed to a su- pervised classifier. With respect to task 3, we develop a naive approach, collecting a set of representations to identify the predefined aspects requested by the organisers. Finally, task 4 uses heuristics to identify the scope of each aspect, to then classify their sentiment via a supervised classifier. The experimental results are promising and will serve us as the starting point to develop more complex techniques., [Resumen]: Este artículo describe nuestra participación en la tercera edición del taller de análisis del sentimiento de tuits escritos en castellano, el tass 2014. En la evaluación competitiva de este año, se han propuesto cuatro retos: (1) análisis del sentimiento a nivel global, (2) clasificación de tópicos, (3) extracción de aspectos y (4) análisis del sentimiento a nivel aspectual. Para las tareas 1 y 2 empleamos una aproximación basada en aprendizaje automático, donde distintos recursos lingüísticos e información extraída del conjunto de entrenamiento son utilizados para entrenar un clasificador supervisado. Para abordar la tarea 3, nuestra aproximación recolecta una lista de representaciones que es empleada para identificar los aspectos requeridos por los organizadores. Por último, la tarea 4 delega en heurísticas para identificar el alcance de cada aspecto, para después determinar su sentimiento a través de un clasificador supervisado. Los resultados experimentales son prometedores y nos servirán para desarrollar técnicas más complejas en el futuro., Research reported in this paper has been partially funded by Ministerio de Educación, Cultura y Deporte (FPU13/01180) and Xunta de Galicia (Grant CN2012/008)., Xunta de Galicia; CN2012/008
RUC. Repositorio da Universidade da Coruña
Publication Contribución de congreso (ConferenceOutput)

LyS at TASS 2015: Deep Learning Experiments for Sentiment Analysis on Spanish Tweets

  • Vilares, David
  • Doval, Yerai
  • Alonso, Miguel A.
  • Gómez-Rodríguez, Carlos
[Abstract]: This paper describes the participation of the LyS group at tass 2015. In this year’s edition, we used a long short-term memory neural network to address the two proposed challenges: (1) sentiment analysis at a global level and (2) aspect-based sentiment analysis on football and political tweets. The performance of this deep learning approach is compared to our last-year model, based on a square-regularized logistic regression. Experimental results show that strategies such as unsupervised pre-training, sentiment-specific word embedding or modifying the current architec- ture might be needed to achieve state-of-the-art results., [Resumen]: Este art´ıculo describe la participaci´on del grupo LyS en el tass 2015. En la edici´on de este a˜no, hemos utilizado una red neuronal denominada long short- term memory para abordar los dos retos propuestos: (1) an´alisis del sentimiento a nivel global y (2) an´alisis del sentimiento a nivel de aspectos sobre tuits futbol´ısticos y de pol´ıtica. El rendimiento obtenido por esta red de aprendizaje profundo es comparado con el de nuestro sistema del a˜no pasado, una regresi´on log´ıstica con una regularizaci´on cuadr´atica. Los resultados experimentales muestran que es necesario incluir estrategias como pre-entrenamiento no supervisado, t´ecnicas espec´ıficas para representar palabras como vectores o modificar la arquitectura actual para alcanzar resultados acordes con el estado del arte., This research is supported by the Ministerio de Economía y Competitividad y FEDER (FFI2014-51978-C2) and Xunta de Galicia (R2014/034). The first author is funded by the Ministerio de Educación, Cultura y Deporte (FPU13/01180)., Xunta de Galicia; R2014/034
RUC. Repositorio da Universidade da Coruña
Publication Contribución de congreso (ConferenceOutput)

Identificación Automática del Idioma en Twitter: Adaptación de Identificadores del Estado del Arte al Contexto Ibérico, Automatic Language Identification in Twitter: Adapting State-of-the-Art Identifiers to the Iberian Context

  • Doval, Yerai
  • Vilares, David
  • Vilares, Jesús
[Abstract]: We describe here our partipation in TweetLID. After having studied the problem of language identification, the resources available, and designed a text conflation approach for this kind of tasks, we joined the competition with two sys- tems: the first one was based in the guesser langdetect, re-trained and adapted in order to work with conflated text; the second one was an approach based on majo- rity vote which used a set of re-trained and adapted classifiers. Results are analized both globally and at language and tweet-type levels., [Resumen]: Describimos aquí nuestra participación en el TweetLID. Tras estudiar la problemática de la identificación del idioma, los recursos disponibles y diseñar una solución para la normalización del texto en este tipo de tareas, presentamos dos sis- temas a competición: el primero basado en el clasificador langdetect, re-entrenado y adaptado para trabajar con texto normalizado; el segundo, una solución basada en la votación de clasificadores individuales igualmente re-entrenados y adaptados. Los resultados son analizados tanto globalmente como por idioma y tipo de tuit., Trabajo parcialmente financiado por el Ministeriode Economía y Competitividad y FEDER (TIN2010-18552-C03-02), Ministerio de Educación, Cultura y Deporte (Beca FPU ref. 13/01180) y Xunta de Galicia (CN2012/008)., Xunta de Galicia; CN2012/008
RUC. Repositorio da Universidade da Coruña
Publication Contribución de congreso (ConferenceOutput)

Detecting Perspectives in Political Debates

  • Vilares, David
  • He, Yulan
[Abstract]: We explore how to detect people’s perspectives that occupy a certain proposition. We propose a Bayesian modelling approach where topics (or propositions) and their associated perspectives (or viewpoints) are modeled as latent variables. Words associated with topics or perspectives follow different generative routes. Based on the extracted perspectives, we can extract the top associated sentences from text to generate a succinct summary which allows a quick glimpse of the main viewpoints in a document. The model is evaluated on debates from the House of Commons of the UK Parliament, revealing perspectives from the debates without the use of labelled data and obtaining better results than previous related solutions under a variety of evaluations., We thank Charles Marshall for crawling the HCD data. DV was funded by MECD (FPU 13/01180), MINECO (FFI2014-51978-C2-2-R) and Inditex-UDC grants for research stays. YH is partly funded by the Natural Science Foundation of China (61528302)., China. Natural Science Foundation of China; 61528302
RUC. Repositorio da Universidade da Coruña
Publication Contribución de congreso (ConferenceOutput)

Seguimiento y análisis automático de contenidos en redes sociales

  • Alonso, Miguel A.
  • Gómez-Rodríguez, Carlos
  • Vilares, David
  • Doval, Yerai
  • Vilares, Jesús
[Abstract]: La Minería de Opiniones es la disciplina que aborda el tratamiento automático de las opiniones contenidas en un texto. Permite, por ejemplo, determinar si en un texto se está opinando o no, o si la polaridad o sentimiento que se expresa en el mismo es positiva, negativa o mixta. También permite la extracción automática de características, lo que posibilita conocer la percepción que los autores tienen sobre aspectos concretos de un tema determinado. Este trabajo, tras realizar una introducción a dicho ámbito, presenta una aproximación propia al mismo, la cual destaca por emplear información sintáctica, así como por estar especialmente adaptada a uno de los contextos de trabajo más complicados, Twitter. Dicha tecnología es fácilmente aplicable a tareas de inteligencia., Investigación financiada por Ministerio de Economía y Competitividad (FFI2014-51978-C2-2-R), Xunta de Galicia (R2014/034) y Ministerio de Educación, Cultura y Deporte (FPU13/01180)., Xunta de Galicia; R2014/034
RUC. Repositorio da Universidade da Coruña
Publication Contribución de congreso (ConferenceOutput)

A non-projective greedy dependency parser with bidirectional LSTMs

  • Vilares, David
  • Gómez-Rodríguez, Carlos
[Abstract]: The LyS-FASTPARSE team present BIST-COVINGTON, a neural implementation of the Covington (2001) algorithm for non-projective dependency parsing. The bidirectional LSTM approach by Kiperwasser and Goldberg (2016) is used to train a greedy parser with a dynamic oracle to mitigate error propagation. The model participated in the CoNLL 2017 UD Shared Task. In spite of not using any ensemble methods and using the baseline segmentation and PoS tagging, the parser obtained good results on both macro-average LAS and UAS in the big treebanks category (55 languages), ranking 7th out of 33 teams. In the all treebanks category (LAS and UAS) we ranked 16th and 12th. The gap between the all and big categories is mainly due to the poor performance on four parallel PUD treebanks, suggesting that some ‘suffixed’ treebanks (e.g. Spanish-AnCora) perform poorly on cross-treebank settings, which does not occur with the corresponding ‘unsuffixed’ treebank (e.g. Spanish). By changing that, we obtain the 11th best LAS among all runs (official and unofficial). The code is made available at https://github.com/CoNLL-UD-2017/LyS-FASTPARSE., David Vilares is funded by an FPU Grant 13/01180. Carlos Gómez-Rodríguez has received funding from the European Research Council (ERC), under the European Union’s Horizon 2020 research and innovation programme (FASTPARSE, grant agreement No 714150). Both authors have received funding from the TELEPARES-UDC project from MINECO.
RUC. Repositorio da Universidade da Coruña
Publication Contribución de congreso (ConferenceOutput)

Sentiment Analysis on Monolingual, Multilingual and Code-Switching Twitter Corpora

  • Vilares, David
  • Alonso, Miguel A.
  • Gómez-Rodríguez, Carlos
[Abstract]: We address the problem of performing po- larity classification on Twitter over differ- ent languages, focusing on English and Spanish, comparing three techniques: (1) a monolingual model which knows the language in which the opinion is written, (2) a monolingual model that acts based on the decision provided by a language iden- tification tool and (3) a multilingual model trained on a multilingual dataset that does not need any language recognition step. Results show that multilingual models are even able to outperform the monolingual models on some monolingual sets. We introduce the first code-switching corpus with sentiment labels, showing the robust- ness of a multilingual approach., This research is supported by the Ministerio de Economía y Competitividad (FFI2014-51978-C2) and Xunta de Galicia (R2014/034). The first authoris funded by the Ministerio de Educación, Cultura y Deporte (FPU13/01180)., Xunta de Galicia; R2014/034
RUC. Repositorio da Universidade da Coruña
Publication Contribución de congreso (ConferenceOutput)

LyS at SemEval-2016 Task 4: Exploiting Neural Activation Values for Twitter Sentiment Classification and Quantification

  • Vilares, David
  • Doval, Yerai
  • Alonso, Miguel A.
  • Gómez-Rodríguez, Carlos
[Abstract]: In this paper we describe our deep learning approach for solving both two-, three- and fiveclass tweet polarity classification, and twoand five-class quantification. We first trained a convolutional neural network using pretrained Twitter word embeddings, so that we could extract the hidden activation values from the hidden layers once some input had been fed to the network. These values were then used as features for a support vector machine in both the classification and quantification subtasks, together with additional linguistic information in the former scenario. The results obtained for the classification subtasks show that this approach performs better than a single convolutional network, and for the quantification part it also yields good results. Official rankings locate us: 2nd (practically tied with 1st) for the binary classification task, 2nd for binary quantification and 4th (practically tied with 3rd) for the five-class polarity classification challenge., This research is supported by the Ministerio de Economía y Competitividad (FFI2014-51978-C2). David Vilares is funded by the Ministerio de Educación, Cultura y Deporte (FPU13/01180). Yerai Doval is funded by the Ministerio de Economía y Competitividad (BES-2015-073768). Carlos Gómez-Rodríguez is funded by an Oportunius program grant (Xunta de Galicia)