Arabic Sentiment Analysis of Eateries’ Reviews: Qassim region Case study

Social media plays an essential role in daily life. It allows people to express their thoughts and feelings about available products on e-commerce websites, which is often called an opinion or review. The aim is to see the author’s mood or the speaker’s attitude, which can be positive or negative about a product. This expression by the people is called sentiment. Sentiment analysis is a machine learning method in which the machine learns and analyzes some textual data’s sentiment and emotions (for example, reviews about movies or products). In this research, a real-world data set was collected from customers in the Qassim region using a Microsoft Form survey, containing approximately 1,785 Arabic reviews related to the Qassim region, Saudi Arabia. The overall aim is to find people’s opinions regarding different eateries (cafes and restaurants) using sentiment analysis of the customers’ reviews written in the Arabic language. The experiments are run using 10-fold cross-validation. The performance of the algorithms was measured using accuracy, recall, and F-measure. Different supervised machine learning classifiers were applied. Support Vector Machine, Logistic regression, and Random Forest achieved the best results. Furthermore, we analyzed the differences between all results in this study.