Implementation Of Convolutional Neural Network (CNN) To Detect Hate Speech And Emotions On Twitter

The research has successfully developed a highly accurate hate speech detection model on CNN-based Twitter. The model focuses on hate speech loaded with negative sentiment related to sensitive issues such as race, religion, and sexual orientation in Indonesian. The research process includes data collection, text pre-processing, and the use of Word2Vec for word representation. Specially designed CNN models are then trained on the datasets. The results show excellent accuracy, which is 87% for emotional assessment and 99% for detection of hate speech. The advantage of this model lies in its ability to capture subtle patterns in language that indicate hate speech. Despite this, the research still has some limitations, such as the limited size of the datasets. Further research is needed to overcome these constraints and improve the performance of the model.