<![CDATA[Hate speech is one of the topics that often discussed in information technology. Hate speech has been usually used by the people that don't like or hate with someone or a group. People stated their hate speech with post it in social media. One of the most used social media to spread the hate speech is Twitter. Hate speech identification is needed to decrease the spread of hate speech. The method used in this research is Naive Bayes based on N-gram and feature selection Information Gain. N-gram features that used in this research are Unigram, Bigram, and combination unigram-bigram. 250 data are used in this research with hate speech label and 250 data with non hate speech label and have 80% proportion for data training and 20% for data testing. The best accuracy results in this research come from Unigram feature and without feature selection Information Gain. The best accuracy result is 84%, precision value 92%, recall value 79,31%, and f-measure value 85,18%. Based on the results obtained it can be concluded that to classify hate speech in Twitter using Naive Bayes has the best result with Unigram feature and without using feature selection Information Gain.]]>
Copyrights © 2019
