SentiWordSKOS: A lexical ontology extended with sentiments and emotions

With the Web 2.0, the user is become an active part of the global network, by interacting with other people through social networks, blogs, etc. and posting his own comments. The importance of user opinions and relative sentiments and emotions is undiscussed. Companies and organizations monitor blog sites to catch the consumers´ opinions about their products and services and then improve their marketing and business plans. In an effort to explore potential resolutions to this issue, the recent research proceeds towards the “distillation” of the social knowledge, in order to capture the moods that govern the users´ behavior on the Web. In the light of these observations, this paper presents SentiWordSKOS, a lightweight ontology which provides a concept-based representation of sentiments and emotions. SentiWordSKOS integrates data from WordNet (describing natural language words) and SentiWordNet (associating a polarity to words) in order to get a mapping between words and sentiments. In addition, each word sense is further associated with a set of the emotions, according to the “Hourglass of Emotions”, a sentic levels based model, used to describe emotions.