A comparison of classification algorithms for hate speech detection

Abstract

Freedom of opinion through social media is frequently affect a negative impact that spreads hatred. This study aims to automatically detect Indonesian tweets that contain hate speech on Twitter social media. The data used amounted to 4,002 tweets related to politics, religion, ethnicity and race in Indonesia. The application model uses classification methods with machine learning algorithms such as Naïve Bayes, Multi Level Perceptron, AdaBoost Classifier, Decision Tree and Support Vector Machine. The study also compared the performance of the model using SMOTE to overcome imbalanced data. The results show that the Multinomial Naive Bayes algorithm produces the best model with the highest recall value of 93.2% which has an accuracy value of 71.2% for the classification of hate speech. Therefore, the Multinomial Naïve Bayes algorithm without SMOTE is recommended as the model to detect hate speech on social media.