Facebook Users Relationships Analysis Based on Sentiment Classification

摘要: It is presented an approach aimed at analyzing the homepage of a Facebook user or group in order to automatically detect who has discussed what and how it been discussed. All public posts shared by are retrieved ad hoc built crawler. Information such as text messages, comments, likes, extracted for each post. Each post classified belonging set predefined categories its sentiment also detected being positive, negative neutral. comments that therefore analyzed categorized together with polarity. For category created graph where highlighted concordance between related comments. The can be used profile relationships according classification.