Supervisors info:
Κωνσταντίνος Μουρλάς, Επίκουρος καθηγητής, Τμήμα Επικοινωνίας και ΜΜΕ, ΕΚΠΑ.
Original Title:
«Ανάλυση πολιτικού λόγου σε μεγάλα σώματα κειμένων και σχολίων στο Twitter, με χρήση τεχνικών εξόρυξης πληροφορίας (text mining) και ανάλυσης φυσικής γλώσσας (NLP): Η μελέτη του φαινομένου της ρητορικής μίσους»
Summary:
The purpose of this thesis is the detection of hate speech in the political and social environment of Greece. The topic of this research is the automatic analysis of journalistic, political discourse, and finally of users’ tweets in a big body of text. Hate Speech, even though it has recently investigated, is profoundly predated, from the 60s. The new technological era has given place to hate speech to rise, in the area of the internet. Furthermore, the imprint of this phenomenon has appeared, because of the need for deliberation of articles, comments, even though images and videos through the internet.
Meanwhile, techniques, such as text mining and natural language processing, which are many times automatic, have given new potentials to social research. Anymore, the researchers can use automatic tools and algorithms to scrutinize topics that matter to society, such as Hate speech. An essential requirement for content analysis - the method of this research - was the overview of the bibliography, but also a survey of other researchers over the field. In this way, methods for collection, processing, analysis, and outcomes from the texts of interest, have been comprehensible.
Specifically, this research has been focused on the fingertip presence of Golden Dawn, one of the extremely right-wing parties, on the internet for the last few years. Moreover, different websites and online newspapers have concerned this research, while in these, there are some articles and comments that have entailed hate speech and extremist views. Likewise, the interest has focused on the presence of hate speech in the political and the social frame. Therefore, the discourse of the plenary session and the tweets of users have been collected and studied.
The results were extremely interesting. Hate speech has existed in the texts of the political discourses of parliament, in a great amount, as well as in the comments of Twitter. Sites that are interested in analyzing political and social views, have encapsulated a huge percentage of hate speech. Despite the efforts to reduce the presence of hate speech, the phenomenon has a strong appearance in everyday discourse.
Keywords:
Hate speech, discourse of hate, social media, twitter, official political discourse, nationalism, Golden Dawn, Twitter, text mining, natural language processing, automatic tools, algorithms.
