Xiaodan Zhu from Canada's National Research Council will speak on the sentiment analysis of social media texts.
Automatically detecting sentiment of product reviews, blogs, tweets, and SMS messages has attracted extensive interest from both the academia and industry. It has a number of applications, including: tracking sentiment towards products, movies, politicians, etc.; improving customer relation models; detecting happiness and well-being; and improving automatic dialogue systems.
This talk will begin with an introduction to sentiment analysis and its various forms on the term level, message level, document level, and aspect level. We will then describe in detail the NRC-Canada's approaches, which yielded the overall best performing models in recent SemEval competitions. We will discuss features that had the most impact. Finally, the talk will flesh out limitations of current approaches and promising future directions.