With the increasing availability of blogs, Internet forums and social networks,
electronic press sites, people have the chance to express their opinions and
sentiments and make them available to everyone. These opinions provide valuable
information for decision-making processes. Therefore, the computational
treatment of sentiments and opinions bearing text has been viewed as a promising,
while challenging, area of research that can serve different purposes. To fulfill
this, we need to identify first named entities, references to pronouns and other
anaphors, and finally decide which of them is the holder of the expressed opinion.