Sentiment Classification of Social Media Content with Features Generated Using Topic Models

Research output: Chapter in Book/Report/Conference proceedingChapter

2 Citations (Scopus)

Abstract

This paper presents a method for using topic distributions generated from topic models as features for performing sentiment analysis on documents. This will be tested in the social media domain, specifically Twitter. The proposed approach allows for the mapping from word space to topic space which allows for less fea- tures to be needed and also reduces computational complexity. Multiple machine learning algorithms will be used to test the topic model generated features and a number of different versions of test corpus will be used, including unigrams, bi- grams, part-of-speech tagging and adjectives only. The method proposed will also be compared to other notable topic-sentiment methods such as the aspect-sentiment unification model and the joint sentiment/topic model. The results show that using topic distributions can improve the accuracy of classification algorithms, however, the performance can be dependent on the algorithm used and the initial features used. Additionally, we show that using only topics as features outperforms the hy- brid topic-sentiment models.
LanguageEnglish
Title of host publicationProceedings of the Eighth European Starting AI Researcher Symposium (STAIRS-2016)
Place of PublicationAmsterdam
PublisherIOS Press
Pages155-165
Volume284
ISBN (Print)978-1-61499-681-1
DOIs
Publication statusPublished - 30 Aug 2016

Fingerprint

Learning algorithms
Learning systems
Computational complexity

Keywords

  • sentiment classification
  • topic models
  • social media
  • feature generation

Cite this

Blair, S., Bi, Y., & Mulvenna, M. (2016). Sentiment Classification of Social Media Content with Features Generated Using Topic Models. In Proceedings of the Eighth European Starting AI Researcher Symposium (STAIRS-2016) (Vol. 284, pp. 155-165). Amsterdam: IOS Press. https://doi.org/10.3233/978-1-61499-682-8-155
Blair, Stuart ; Bi, Yaxin ; Mulvenna, Maurice. / Sentiment Classification of Social Media Content with Features Generated Using Topic Models. Proceedings of the Eighth European Starting AI Researcher Symposium (STAIRS-2016). Vol. 284 Amsterdam : IOS Press, 2016. pp. 155-165
@inbook{8c99ed9186c444489940d90020fc855e,
title = "Sentiment Classification of Social Media Content with Features Generated Using Topic Models",
abstract = "This paper presents a method for using topic distributions generated from topic models as features for performing sentiment analysis on documents. This will be tested in the social media domain, specifically Twitter. The proposed approach allows for the mapping from word space to topic space which allows for less fea- tures to be needed and also reduces computational complexity. Multiple machine learning algorithms will be used to test the topic model generated features and a number of different versions of test corpus will be used, including unigrams, bi- grams, part-of-speech tagging and adjectives only. The method proposed will also be compared to other notable topic-sentiment methods such as the aspect-sentiment unification model and the joint sentiment/topic model. The results show that using topic distributions can improve the accuracy of classification algorithms, however, the performance can be dependent on the algorithm used and the initial features used. Additionally, we show that using only topics as features outperforms the hy- brid topic-sentiment models.",
keywords = "sentiment classification, topic models, social media, feature generation",
author = "Stuart Blair and Yaxin Bi and Maurice Mulvenna",
year = "2016",
month = "8",
day = "30",
doi = "10.3233/978-1-61499-682-8-155",
language = "English",
isbn = "978-1-61499-681-1",
volume = "284",
pages = "155--165",
booktitle = "Proceedings of the Eighth European Starting AI Researcher Symposium (STAIRS-2016)",
publisher = "IOS Press",
address = "Netherlands",

}

Blair, S, Bi, Y & Mulvenna, M 2016, Sentiment Classification of Social Media Content with Features Generated Using Topic Models. in Proceedings of the Eighth European Starting AI Researcher Symposium (STAIRS-2016). vol. 284, IOS Press, Amsterdam, pp. 155-165. https://doi.org/10.3233/978-1-61499-682-8-155

Sentiment Classification of Social Media Content with Features Generated Using Topic Models. / Blair, Stuart; Bi, Yaxin; Mulvenna, Maurice.

Proceedings of the Eighth European Starting AI Researcher Symposium (STAIRS-2016). Vol. 284 Amsterdam : IOS Press, 2016. p. 155-165.

Research output: Chapter in Book/Report/Conference proceedingChapter

TY - CHAP

T1 - Sentiment Classification of Social Media Content with Features Generated Using Topic Models

AU - Blair, Stuart

AU - Bi, Yaxin

AU - Mulvenna, Maurice

PY - 2016/8/30

Y1 - 2016/8/30

N2 - This paper presents a method for using topic distributions generated from topic models as features for performing sentiment analysis on documents. This will be tested in the social media domain, specifically Twitter. The proposed approach allows for the mapping from word space to topic space which allows for less fea- tures to be needed and also reduces computational complexity. Multiple machine learning algorithms will be used to test the topic model generated features and a number of different versions of test corpus will be used, including unigrams, bi- grams, part-of-speech tagging and adjectives only. The method proposed will also be compared to other notable topic-sentiment methods such as the aspect-sentiment unification model and the joint sentiment/topic model. The results show that using topic distributions can improve the accuracy of classification algorithms, however, the performance can be dependent on the algorithm used and the initial features used. Additionally, we show that using only topics as features outperforms the hy- brid topic-sentiment models.

AB - This paper presents a method for using topic distributions generated from topic models as features for performing sentiment analysis on documents. This will be tested in the social media domain, specifically Twitter. The proposed approach allows for the mapping from word space to topic space which allows for less fea- tures to be needed and also reduces computational complexity. Multiple machine learning algorithms will be used to test the topic model generated features and a number of different versions of test corpus will be used, including unigrams, bi- grams, part-of-speech tagging and adjectives only. The method proposed will also be compared to other notable topic-sentiment methods such as the aspect-sentiment unification model and the joint sentiment/topic model. The results show that using topic distributions can improve the accuracy of classification algorithms, however, the performance can be dependent on the algorithm used and the initial features used. Additionally, we show that using only topics as features outperforms the hy- brid topic-sentiment models.

KW - sentiment classification

KW - topic models

KW - social media

KW - feature generation

U2 - 10.3233/978-1-61499-682-8-155

DO - 10.3233/978-1-61499-682-8-155

M3 - Chapter

SN - 978-1-61499-681-1

VL - 284

SP - 155

EP - 165

BT - Proceedings of the Eighth European Starting AI Researcher Symposium (STAIRS-2016)

PB - IOS Press

CY - Amsterdam

ER -

Blair S, Bi Y, Mulvenna M. Sentiment Classification of Social Media Content with Features Generated Using Topic Models. In Proceedings of the Eighth European Starting AI Researcher Symposium (STAIRS-2016). Vol. 284. Amsterdam: IOS Press. 2016. p. 155-165 https://doi.org/10.3233/978-1-61499-682-8-155