Ontology-based sentiment analysis and community detection on social media: application to Brexit

Sentiment Analysis and Community Detection are two of the main methods used to analyze and comprehend human interactions on social media. These domains expanded immensely with the rise of social media, as it provided a free and ever-increasing quantity of data. Domain ontologies are of great assistance in collecting specific data, as it describes the domain's features and their existing relationships. Therefore, we utilize them in collecting subject-specific data on social media. This paper describes the framework we've designed in order to understand, in depth, the impact of a subject on social media users, and also to evaluate the difference between the Lexicon Approach and the Machine Learning Approach, by assessing the strengths and weaknesses of each. This framework also aims to deeply understand the connections that exist between users, depending on their point of view on a particular subject. The resulting framework not only analyzes textual data (by taking into account the negation and sentence POS tags), but also visual one, such as images. In order to test the framework, we chose to analyze the Brexit phenomenon by collecting ontology-based data from Twitter and Reddit, and it had some promising results.