Browse DORAS
Browse Theses
Latest Additions
Creative Commons License
Except where otherwise noted, content on this site is licensed for use under a:

Sentiment analysis and real-time microblog search

Bermingham, Adam (2012) Sentiment analysis and real-time microblog search. PhD thesis, Dublin City University.

Full text available as:

PDF - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader


This thesis sets out to examine the role played by sentiment in real-time microblog search. The recent prominence of the real-time web is proving both challenging and disruptive for a number of areas of research, notably information retrieval and web data mining. User-generated content on the real-time web is perhaps best epitomised by content on microblogging platforms, such as Twitter. Given the substantial quantity of microblog posts that may be relevant to a user query at a given point in time, automated methods are required to enable users to sift through this information. As an area of research reaching maturity, sentiment analysis offers a promising direction for modelling the text content in microblog streams. In this thesis we review the real-time web as a new area of focus for sentiment analysis, with a specific focus on microblogging. We propose a system and method for evaluating the effect of sentiment on perceived search quality in real-time microblog search scenarios. Initially we provide an evaluation of sentiment analysis using supervised learning for classi- fying the short, informal content in microblog posts. We then evaluate our sentiment-based filtering system for microblog search in a user study with simulated real-time scenarios. Lastly, we conduct real-time user studies for the live broadcast of the popular television programme, the X Factor, and for the Leaders Debate during the Irish General Election. We find that we are able to satisfactorily classify positive, negative and neutral sentiment in microblog posts. We also find a significant role played by sentiment in many microblog search scenarios, observing some detrimental effects in filtering out certain sentiment types. We make a series of observations regarding associations between document-level sentiment and user feedback, including associations with user profile attributes, and users’ prior topic sentiment.

Item Type:Thesis (PhD)
Date of Award:March 2012
Additional Information:Winner of Irish Software Association project with most commercial potential, 2011.
Supervisor(s):Smeaton, Alan F.
Uncontrolled Keywords:Sentiment; social networks; Twitter; Microblogs;
Subjects:Computer Science > Interactive computer systems
Computer Science > Computational linguistics
Computer Science > Machine learning
Computer Science > Information retrieval
Computer Science > World Wide Web
DCU Faculties and Centres:DCU Faculties and Schools > Faculty of Engineering and Computing > School of Computing
Research Initiatives and Centres > CLARITY: The Centre for Sensor Web Technologies
Use License:This item is licensed under a Creative Commons Attribution-NonCommercial-No Derivative Works 3.0 License. View License
Funders:Science Foundation Ireland
ID Code:16748
Deposited On:28 Mar 2012 14:02 by Alan Smeaton. Last Modified 15 Jun 2012 09:47

Download statistics

Archive Staff Only: edit this record