Enhanced models for arabic hate speech sentiment analysis: A comparative study using SVM, CNN, and Naive Bayes

The aim of this paper is to develop effective machine learning and deep learning models using genetic algorithms that can accurately classify hate speech sentiments and contribute to mitigating the negative impact of hate speech online. The research explores different techniques for data preprocessing, feature selection, and model training, including genetic support vector machines (SVM), genetic naive Bayes algorithm, and genetic convolutional neural networks (CNN). Performance evaluation is performed using metrics such as precision, recall, F1 score, and accuracy. When comparing the previous classifiers, it was observed that the SVM algorithm was superior with an accuracy of 0.93 and an F1 score of 0.96, followed by the deep learning genetic networks with an accuracy of 0.91 and an F1 score of 0.92, and finally the naive genetic Bayes algorithm with an accuracy of 0.87 and an F1 score of 0.92. The paper presented a comparison between its results and the results of previous studies conducted on the same dataset, and it was observed that the SVM algorithm was superior with an accuracy of 0.93 over the study that used AraBERT, which gave an accuracy of 0.80, and over the study that used SVM, which gave an accuracy of 0.87.