Polarity Classification of Tweets Considering the Poster’s Emotional Change by a Combination of Naive Bayes and LSTM

Kiichi Tago*, Kosuke Takagi, Qun Jin

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Twitter, as a popular social networking service, is used all over the world, with which users post tweets for various purposes. When users post tweets, an emotion may be behind the messages. As the emotion changes over time, we should better consider their emotional changes and states when analyzing the tweets. In this study, we improve polarity classification by considering the poster’s emotional state. Firstly, we analyze the sentence structure of a tweet and calculate emotion scores for each category by Naive Bayes. Then, the poster’s emotion state is estimated by the emotion scores, and a prediction model of emotional state is created by Long Short Term Memory (LSTM). Based on the predicted emotional state, weights are added to the scores. Finally, polarity classification is performed based on the weighted emotion scores for each category. In our experiments, our approach showed better accuracy than other related studies.

Original languageEnglish
Title of host publicationComputational Science and Its Applications – ICCSA 2019 - 19th International Conference, 2019, Proceedings
EditorsSanjay Misra, Elena Stankova, Vladimir Korkhov, Carmelo Torre, Eufemia Tarantino, Ana Maria A.C. Rocha, David Taniar, Osvaldo Gervasi, Bernady O. Apduhan, Beniamino Murgante
PublisherSpringer-Verlag
Pages579-588
Number of pages10
ISBN (Print)9783030242886
DOIs
Publication statusPublished - 2019 Jan 1
Event19th International Conference on Computational Science and Its Applications, ICCSA 2019 - Saint Petersburg, Russian Federation
Duration: 2019 Jul 12019 Jul 4

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11619 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference19th International Conference on Computational Science and Its Applications, ICCSA 2019
Country/TerritoryRussian Federation
CitySaint Petersburg
Period19/7/119/7/4

Keywords

  • Deep learning
  • Naive Bayes
  • Polarity classification
  • Twitter

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint

Dive into the research topics of 'Polarity Classification of Tweets Considering the Poster’s Emotional Change by a Combination of Naive Bayes and LSTM'. Together they form a unique fingerprint.

Cite this