Utilizing Deep Learning in Arabic Text Classification Sentiment Analysis of Twitter

The number of social media users has increased. These share and reshare their ideas in posts this information can be mined used by decision-makers different domains, who analyse study user opinions on networks to improve the quality products or specific phenomena. During COVID-19 pandemic, was make decisions limit spread disease using sentiment analysis. Substantial research topic been done; however, there are limited Arabic textual resources media. This resulted fewer analyses texts. proposes a model for analysis Twitter dataset deep learning models with word embedding. It uses supervised algorithms proposed dataset. contains 51,000 tweets, which 8,820 classified as positive, 37,360 neutral, negative. After cleaning it will contain 31,413. experiment carried out applying models, Convolutional Neural Network Long Short-Term Memory while comparing results machine techniques such Naive Bayes Support Vector Machine. accuracy AraBERT is 0.92% when test 3,505 tweets.