PELESent: Cross-Domain Polarity Classification Using Distant Supervision

Corrêa, Edilson A.; Marinho, Vanessa Queiroz; Santos, Leandro Borges dos; Bertaglia, Thales Felipe Costa; Treviso, Marcos; Brum, Henrico Bertini

doi:10.1109/bracis.2017.45

Cited by 12 publications

(5 citation statements)

References 20 publications

(44 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Portuguese is the lack of standard annotated datasets. For the accomplishment of the present work we have considered some of the few recent corpus of tweets in Brazilian Portuguese cited in the literature, including the PELESent [3], Tweet-SentBR [4] and BRTweetSentCorpus [5].…”

Section: Datasets For Brazilian Portuguese Sentiment Analysismentioning

confidence: 99%

Sentiment Analysis of Tweets in Brazilian Portuguese with Convolutional Neural Networks

Coello

Neto

2019

Int J Innov Educ Res

View full text Add to dashboard Cite

Sentiment analysis of texts posted on Twitter is a natural language processing task whose importance has grown along with the increase in the number of users of the platform and the interest of organizations on the opinions of their employees, customers and users.Although Brazil is the sixth country in the world with most active users of Tweeter and Portuguese is the seventh most spoken language in the world, with 221 million speakers (200 million of them living in Brazil), the number of articles that discuss sentiment analysis approaches for Brazilian Portuguese is a small fraction of those that focus on the English language. On the other hand, few works use deep learning for this task when compared with other machine learning and lexical based methods. In this context, the work described in this article addresses the problem using Convolutional Neural Networks (CNN). The paper presents the results of an experimental evaluation that shows that a CNN with a relatively simple architecture can perform much better than a previous approach that uses ensembles of other machine learning classifiers combined with text preprocessing heuristics

show abstract

Section: Datasets For Brazilian Portuguese Sentiment Analysismentioning

confidence: 99%

Sentiment Analysis of Tweets in Brazilian Portuguese with Convolutional Neural Networks

Coello

Neto

2019

Int J Innov Educ Res

View full text Add to dashboard Cite

show abstract

“…Mais detalhes sobre seu funcionamento podem ser encontrados na Subseção 3.2.3. Esse tipo de arquitetura já foi utilizado com sucesso para segmentação de sentenças (TILK; ALUMÄE, 2016;CHE et al, 2016) e também para classificação de textos em geral (LAI et al, 2015), inclusive em trabalhos para o português brasileiro, como em (TREVISO; SHULBY; ALUÍSIO, 2017;CORREA et al, 2017). A arquitetura da RCNN utilizada, baseada no trabalho de Treviso, Shulby e Aluísio (2017) utilizado, sendo que o processo de indução das word embeddings usadas é descrito detalhadamente na Seção 6.3.…”

Section: Identificação De Ruídos Baseada Em Aprendizado De Máquinaunclassified

Normalização textual de conteúdo gerado por usuário

Bertaglia¹

View full text Add to dashboard Cite

User Generated Content (UGC) is the name given to content created spontaneously by ordinary individuals, without connections to the media. This type of content carries valuable information and can be exploited by several areas of knowledge. Much of the UGC is provided in the form of texts -product reviews, comments on forums about movies, and discussions on social networks are examples. However, the language used in UGC texts differs, in many ways, from the cultured norm of the language, making it difficult for NLP techniques to handle them. UGC language is strongly linked to the language used in daily life, containing a large amount of noise. Spelling mistakes, abbreviations, slang, absence or misuse of punctuation and capitalization are some noises that make it difficult to process these texts. Several works report considerable loss of performance when testing NLP state-of-the-art tools in UGC texts. Textual Normalization is the process of turning noisy words into words considered correct and can be used to improve the quality of UGC texts. This work reports the development of methods and systems that aim to (a) identify noisy words in UGC, (b) find candidate words for substitution, and (c) rank candidates for normalization. For the identification of noisy words, lexical-based methods and machine learning ones using deep neural networks were proposed. The automatic identification presented results comparable to the use of lexicons, proving that this process can be done with low dependence of resources. For the generation and ranking of candidates, techniques based on lexical similarity and word embeddings were investigated. It was concluded that the use of embeddings is highly suitable for normalization, having achieved the best results. All proposed methods were evaluated based on a UGC corpus annotated throughout the project, containing texts from different sources: discussion forums, product reviews and tweets. A system, Enelvo, combining all methods was implemented and compared to another existing normalizing system, UGCNormal. The results obtained by the Enelvo system were considerably higher, with a correction rate between 67 % and 97 % for different types of noise, with less dependence on resources and greater flexibility in normalization.

show abstract

“…Distant supervision or weak supervision increases the data size (unlabeled text) by (semi-) automatically assigning labels from an external source or using some heuristics. For example , Corrêa Jr et al (2017) applied distant supervision for increasing data size in a twitter-based sentiment analysis corpus. The authors used emojis like ":)" and ":(" for annotating tweets as positive or negative, respectively.…”

Section: Additional Labeled Data Generationmentioning

confidence: 99%

Geração de linguagem natural por meio de representações semânticas abstratas para o português do Brasil

Cabezudo¹

View full text Add to dashboard Cite

This work is dedicated to my lovely family, mi wife Katherine Vasquez, my daughter Natalia Sobrevilla, my mother Angela Cabezudo, my sister Mayra Sobrevilla, my mamita Victoria Carrizales, and my uncle Cesar Cabezudo. Besides, I dedicate this work to all future PhD students who feel they have lost their way along their own research. You will arrive at the correct place, be patient and keep researching!To my life partner, my wife and my best friend, Katherine Vasquez, who supported me unconditionally all this time. You resisted so many work nights and times in which, despite being present, I was a bit absent. I also would like to thank my beautiful daughter, Natalia, you are the best thing that happened to me, and you have been my impulse to give the best of me in this last period.To my family who, despite the distance, were always with me encouraging me on this long walk (and asking every day: "Are you going to finish your thesis soon?"). To my mom Angela Cabezudo, my sister Mayra Sobrevilla, my "mamita" Victoria Carrizales, my uncle Cesar Cabezudo, my dad Mauro Sobrevilla and Iván Ferrer. You are the best family I could have had.To my advisor, prof. Dr. Thiago Alexandre Salgueiro Pardo, for the guidance, the learning acquired, his good cheer and patience. For cheering me up when I felt lost and accepting me again as his student after the master :)To my friends from the Methodist Church in São Carlos because they have been my family during these years. I want to thank the pastors, Levi Pereira, Gilson Sales and Nuria Lisboa and the members of the church :). In particular, I would also like to thank José Mendoza,

show abstract

PELESent: Cross-Domain Polarity Classification Using Distant Supervision

Cited by 12 publications

References 20 publications

Sentiment Analysis of Tweets in Brazilian Portuguese with Convolutional Neural Networks

Sentiment Analysis of Tweets in Brazilian Portuguese with Convolutional Neural Networks

Normalização textual de conteúdo gerado por usuário

Geração de linguagem natural por meio de representações semânticas abstratas para o português do Brasil

Contact Info

Product

Resources

About