Abstract

AbstractHate speech and cyber-harassment have been a major concern on the Internet for a long time. Furthermore, social media platforms, particularly Twitter and Facebook, have elevated it to a worldwide platform on which hate speeches can spread much faster. Manually detecting hate speech from social media is a time-consuming process. Hence, several studies are still being conducted in this field. We build a two-layer hybrid machine learning model using existing machine learning algorithms. This hybrid machine learning algorithm is capable of efficiently detecting hate speech from social media texts. The hybrid approach combines nine different machine learning algorithms to make one hybrid machine learning model. Additionally, we used the bag-of-words and TF-IDF techniques with the two-gram approach to extract the features. Significant experiments are carried out on the hate speech dataset. The accuracy gained by the hybrid machine learning model is much higher than that of available conventional machine learning models.KeywordsNatural language processingHybrid machine learningClassification algorithm

Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call