Abstract:
Social media invites anyone who is interested to participate by contributing and giving
feedback openly, giving comments, and sharing information in a fast and unlimited time.
One of the most widely used and influential social media in Indonesia is Twitter. Twitter
usually contains meaningless chatter, conversation, repeated messages or retweets, selfpromotion,
spam, news and some statements, opinions and expressions. Tweets can
contain blasphemous or hate speech and Good Speech. This study aims to build an
application to classify user tweets against posts on Twitter social media so that it is known
whether the Tweet contains hate speech or not by implementing the multinomial nave
Bayes method. Crawling data using the R programming language using the Rstudio tool.
Use the Twitter API to search data with the keyword "president". The data is taken from
comments containing keywords using a time filter, namely October 2018 - April 2019.
The application is able to classify tweet data using multinomial nave Bayes with an
accuracy of 87,5% based on 400 tweet data by dividing 360 training data and 40 testing
data.