Arabic Sentiment Analysis of Eateries’ Reviews Using Deep Learning
Leen Muteb Alharbi, Ali Mustafa Qamar · Ingénierie des systèmes d information · 2022
After air and water, food is the third most essential thing for humans to provide energy and development. More and more customers of eateries, including restaurants and cafes, express their opinion or sentiment about quality, ambiance, and facilities. This research performs a sentiment analysis of the eateries’ reviews obtained in Qassim, Saudi Arabia. The reviews are obtained in Arabic, the local language of the region. We apply various models of Long Short-Term Memory, a deep learning technique. The best approach achieved 83% accuracy. Furthermore, we also compared the proposed methods with state-of-the-art machine learning ones, such as support vector machines, nearest neighbor, Naïve Bayes, random forest, and logistic regression. The achieved results are promising.