Sentiment Analysis: Subjectivity, Opinion and Point of View
9.30 - 11pm
Michael Strube (EML Research gGmbH Heidelberg)
Companies search the internet for criticism on their products,
politicians want to know if bloggers write bad things about them,
movie freaks look for praise and blame of the latest movie. These are
only a few of the information needs which can be solved by
applications in an emerging research area in Natural Language
Processing: Sentiment Analysis.
In this course we will cover text-level approaches to sentiment analysis as well as sentence-level ones. Relevant NLP tasks are: distinguishing subjective statements from objective ones, determining the perspective of authors, extracting their opinion. We will start with the problem how to acquire annotated data for these tasks. For some tasks they are freely available, for others they need to be created manually. Then we will proceed with elaborating methods for sentiment analysis, recognizing subjectivity and opinion mining.
Bo Pang & Lillian Lee (2008). Opinion Mining and Sentiment Analysis. Foundations and Trends in Information Retrieval, Vol. 2, 135 pages. Now Publishers (the book is available online).