Slides: Context Aware Sentiment Detection

Abstract

The simplicity of using Web publishing services and social networking platforms has resulted in an abundance of user-generated content. A significant portion of this content contains user opinions with clear economic relevance - customer and travel reviews, for example, or the articles of respected bloggers who influence purchase decisions. Analyzing and acting upon user-generated content is therefore becoming imperative for marketers and social scientists who need to gather feedback from very large user communities. In order to identify trends in user-generated content and compare differing perceptions of interest groups, automated sentiment detection identifies and aggregates polar opinions (i.e. positive or negative statements about facts). For achieving accurate results, sentiment detection requires a correct interpretation of natural languages, which remains a challenging task due to their inherent ambiguities. Most approaches to sentiment detection are based on the notion that there is a stable conceptual connection between words and their adjacent text, but neglect the context of opinionated terms when trying to resolve ambiguities. To address this limitation, the presenters will discuss an approach based on contextualized sentiment lexicons and introduce a domain-specific method for the automated refinement of such lexicons.