A Systemic Approach to Appraisal: Identifying Opinion and Sentiment in Text

Опубликовано 6 сентября 2016, 5:12

Sentiment analysis aims to identify the subjective content of text: what kinds of opinions are being expressed, and how? This is a relatively new field with many applications in information extraction and natural language processing.    Systemic Functional Linguistics deals explicitly with how we use language to express sentiment, using a set of systems called Appraisal. This talk introduces a framework for extracting Appraisal, based explicitly on this theory and using new and existing computational techniques.   The talk will cover an introduction to SFL and Appraisal, an explanation of the computational processes involved, and present results from recent experiments in using this approach to classify movie and product reviews.