Hate speech detection from code-mixed Hindi-English tweets using deep learning models. This project reports an increment to the state-of-the-art in hate speech detection for English-Hindi code-mixed tweets. The models result in an improvement of about 12% in F-score over a past work that used statistical classifiers. Details of the project will be soon explaied here.
The trained embeddings are available in the form of a pickle file above.
Satyajit Kamble, Aditya Joshi, `Hate Speech Detection from Code-mixed Hindi-English Tweets Using Deep Learning Models', International Conference on Natural Language Processing 2018, Patiala, India, December 2018.