Identification and Blocking of Hate Speech in English-Hausa Code-mixed Language on Facebook: A Hybrid Deep Learning Framework

The rise of social media platforms, especially Facebook, has revolutionized communication and interaction worldwide.However, the unrestricted nature of these platforms has also given rise to the proliferation of hate speech, posing significant challenges to maintaining a safe and inclusive online environment.In this paper, a novel framework using hybrid deep learning models was proposed for the detection and blocking of hate speech on Facebook English-Hausa code-mixed language.The proposed hybrid deep learning model combines the strengths of convolutional neural networks (CNNs) and recurrent neural networks (RNNs) to form a Convolutional -Long Short-Term Memory (C-LSTM) with glove word embeddings to efficiently identify and block hate speech content.The datasets used in this work were collected from Facebook and were labeled manually as hate speech or non-hate speech.This paper also analyses the model's architecture, data preprocessing techniques, and the evaluation metrics presented, hence demonstrating its effectiveness in real-world scenarios where the experimental results show that the hybrid C-LSTM model generated 0.90 accuracy which outperformed the CNN, LSTM, and the baseline-models.The findings in this paper suggest that hate speech in Facebook English-Hausa code-mixed language can be detected and blocked instantly before posting.