Identifying and Categorising Profane Words in Hate Speech

Abstract

This study attempts to explore the different types of Hate Speech appearing in social media by identifying profane words used in hate speech. This study also compares the profane words used in different generations to assist in identifying the user's profile. Fivehundred (500) comments posted on YouTube on the abusive topics were collected. Profane words are classified into eight different types of hate speech. The finding shows 35% of profane words found in our sample are words related to sexual orientation. Comparison of the terms between 1970 and 2017 also show a high percentage of profane words are sexual orientation. Though the results are found based on only 500 comments collected from YouTube link in the current study, they are useful in establishing the list of profane words which will serve as the base for automatic hate speech identification in our future study. The originality of this research is the development of a training list of profane words for each category and comparison of the type of the words used in 1970 century with today's social media platform.

A Framework for Detecting Intentions of Criminal Acts in Social Media: A Case Study on Twitter

“…Lexical dictionaries and their limitations in feelings evaluation were addressed in [14,55,56]. In Teodorescu et al [55], each slang expression is manually annotated with meanings assigned to it.…”

“…According to Teh et al [14], the use of a lexical dictionary is not sufficient to detect the presence of hate speech in texts written in natural language. The authors indicated that the evolution of vocabulary must be considered for this purpose [14]. Teh et al [14] organized hate terms into eight categories.…”

“…The authors indicated that the evolution of vocabulary must be considered for this purpose [14]. Teh et al [14] organized hate terms into eight categories. A lexical dictionary that correlates a word with a feeling from the Plutchik's wheel of emotions was used in [56].…”

“…According to Agarwal and Sureka [13], only lexical dictionaries are not enough for building automatic intention detection mechanisms. Automated detection using an exclusively lexical approach proved to be flawed [14], requiring the adoption of multiple techniques, such as Machine Learning (ML) ones.…”

Um Estudo Sobre Análise, Representação E Detecção De Intenções De Criminosos Em Postagens Em Mídia Social

“…Segundo Waseem, Thorne, & Bingel (2018), os trabalhos de análise de sentimentos não consideram adequadamente a influência de questões sociais, geográficas e culturais; tal fator prejudica sua eficácia. Segundo Teh et al (2018), a utilização de um dicionário léxico não é suficiente para detectar a presença de discurso de ódio em textos escritos em linguagem natural, uma vez que a evolução do vocabulário deve ser considerada. O estudo propõe a categorização dos termos de ódio em 8 categorias distintas.…”