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

Stuart Blair, Yaxin Bi, Maurice Mulvenna

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review

2 Citations (Scopus)
47 Downloads (Pure)


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.
Original languageEnglish
Title of host publicationProceedings of the Eighth European Starting AI Researcher Symposium (STAIRS-2016)
Place of PublicationAmsterdam
PublisherIOS Press
ISBN (Print)978-1-61499-681-1
Publication statusPublished (in print/issue) - 30 Aug 2016


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


Dive into the research topics of 'Sentiment Classification of Social Media Content with Features Generated Using Topic Models'. Together they form a unique fingerprint.

Cite this