Identifikasi Hate Speech pada Media Sosial Twitter Menggunakan Backpropagation Artificial Neural Network

Abstract

There are many ways to share information nowadays. Twitter is one of the social media that provides many characteristic services. On twitter, many public opinions are expressed in the form of comments. Opinion on a comment can be either positive or negative. The spread of negative information among the public cannot be avoided. User's emotions are not controlled in commenting which makes it easy for other users to be provoked. Currently, hate speech cases are still being studied and concluded manually by the police and experts by reading case documents. However, this method causes the investigation and action against the perpetrators of hate speech to take a long time. Thus, a computational approach is needed that can identify hate speech tweets for time efficiency with proper accuracy. Therefore, identification of hate speech with a fast computational process that is easy to implement with a simple and effective structure is needed at this time. To simplify the identification process, use tweets containing hate speech as data that will later be processed, so that the results can help the authorities. This study uses an Artificial Neural Network (ANN) algorithm with the Backpropogation method. The results of this test obtained values for accuracy of 87.1%, precision of 89.66%, and recall of 83.87% with a total of 1550 datasets.