23 maig 2017
Xavier Ferrer Aran

In this seminar I focus on a particular interesting type of web user-generated content: people’s experiences. We introduce and extend our previous work on aspect extraction and sentiment analysis and propose a novel approach to creating a vocabulary of concepts that we use to characterize a set of entities. The concept vocabulary is created by analyzing the usage of the aspects by the individuals in their textual experiences with entities, and allows us to characterize products with those features with a clear positive and negative polarity to create the bundles of arguments. As I will show, the bundles of arguments allow the reuse of the experiences of other people for other individuals with specific preferences.