Computational Approaches to Subjectivity and Sentiment Analysis

A central point in the quality of communication using natural language is effectiveness, i.e. the extent to which the meaning conveyed through the message reaches the audience and is interpreted in the same manner as it was intended. Rhetorics, as the art of using language to communicate effectively, is based on three pillars: logos (the use of words), ethos (appealing to common concepts) and pathos (appealing to the emotions of the audience). It is thus of no wonder that the fields where communication effectiveness is vital, such as politics, advertisement or marketing, employ in their text the persuasive power of emotions. Affect expression is not only important because of its argumentative power. It also helps to create a more natural, human way of interaction. This is why affect modelling has also become an important research theme in Artificial Intelligence, within the framework of Affective Computing. Research in this area is currently applied to Human-Computer Interaction, intelligent agents modelling and Robotics. Research in automatic Subjectivity and Sentiment Analysis, as subtasks in Affective Computing within Natural Language Processing, has flourished in the past years, as the Social Web made it possible for people all over the world to express, comment or consult opinions on any give