This research discusses the development of a model that can be used to detect hate speech in Indonesia's popular social media. Several methods are collected, analyzed, and implemented to build a framework that is practical and can be used to detect hate speech. The research methodology starts with the study of literature, analysis, and design of the model. From the analysis results, N-Gram, Word2vec, LSTM and emotion classification is used as part of the process of the model. This model consists of ten processes that are carried out sequentially so that comments collected through Facebook and Whatsapp social media can be identified when they contain hate speech. The model also considers the applicable law in Indonesia to facilitate the legal handling of perpetrators.
Copyrights © 2020