Classified hate speech detection in Sinhala tweets

Abstract

"This research explores the use of Natural Language Processing (NLP) and deep learning techniques for hate speech detection in Sinhala tweets with a specific classification. With the increasing prevalence of digital platforms offer potential individuals/groups to use hate speech in them and spread toxicity. This novel approach tries to counter that with the use of 7 Long Short-Term Memory networks. Due to the scarcity of proper datasets the author created data sets manually in order to train all the LSTM models. Each model used a binary classification to identify the specific form of hate speech. An accuracy and F1 score of 90% was achieved by one model which was the highest rated one. This study computed a solution to find classified hate speech in Sinhala text."