NetAC, an automatic classifier of online hate speech comments

Abstract

Nowadays in many linguistic and social areas researchers collect the reaction of people to someone’s statements (newspaper articles or social network posts) with the intention of analyzing the speech style. From that analysis different conclusions can be inferred giving rise to a large number of social impact attitudes. However it is not enough to create a huge corpus of texts. It is necessary to process the collected statements and comments and resort to appropriate tools to extract the relevant terms from the texts and analyze their occurrences. This paper is about a statistical framework, NetAC, built in the context of NetLang Project to study prejudice discourse aiming at individual or group discrimination. Given a categorization table the tools included in NetAC search for frequency of occurrence of the keywords in each category and, based on the greatest frequency, propose a classification for each comment and for the overall text. Besides the main classifier, other features will be presented.