Degree
Master of Science in Computer Science
Department
Department of Computer Science
Faculty / School
Faculty of Computer Sciences (FCS)
Date of Submission
2016-01-01
Supervisor
Dr. Sajjad Haider, Professor, Department of Computer Science
Document type
MSCS Survey Report
Abstract
Sentiment Analysis has become a popular area of research due to exponential increase in the volume of user generated content with the rise of Internet and Social Networking Websites. Sentiment Analysis can be performed at different level of granularities such as document level, sentence level and aspect level. Aspect level sentiment analysis provides more deeper insights on how customer feels about product service. Aspect term extraction, aspect category detection and sentiment polarity detection are the core tasks in aspect based opinion mining. This report presents existing state-of-the-art approaches for performing aspect level opinion mining. We have also provided a comprehensive list of publicly available lexicons and annotated datasets that can be used to build aspect level sentiment analysis systems. The first chapter introduces the research area of Opinion Mining. The second chapter focuses on aspect level opinion mining. In the next three chapters, we have discussed approaches of aspect term extraction, aspect category detection and aspect polarity identification. Taxonomy of various approaches discussed in literature related to opinion mining is presented in Chapter 6. Chapter 7 presents evaluation metrics and various publicly available lexicons and datasets. Chapter 8 provides details of our tool named ABSA Toolkit developed using state-of-art techniques of aspect level opinion mining. The report concludes with some emerging techniques and research issues in opinion mining.
Recommended Citation
Nasim, Z. (2016). Aspect based sentiment analysis (Unpublished MSCS survey report). Institute of Business Administration, Pakistan. Retrieved from https://ir.iba.edu.pk/survey-reports-mscs/135
The full text of this document is only accessible to authorized users.
