AraMAMS: Arabic Multi-Aspect, Multi-Sentiment Restaurants Reviews Corpus for Aspect-Based Sentiment Analysis

The abundance of data on the internet makes analysis a must. Aspect-based sentiment helps extract valuable information from textual data. Because limited Arabic resources, this paper enriches dataset landscape by creating AraMA, first and largest multi-aspect corpus. AraMA comprises 10,750 Google Maps reviews for restaurants in Riyadh, Saudi Arabia. It covers four aspect categories—food, environment, service, price—along with polarities: positive, negative, neutral, conflict. All are labeled at least two categories. A second version, named AraMAMS, includes different sentiments, making it multi-aspect, multi-sentiment dataset. AraMAMS has 5312 covering same categories polarities. Both corpora were evaluated using naïve biased (NB), support vector classification (SVC), linear SVC, stochastic gradient descent (SGD) models. In corpus, task achieved 91.41% F1 measure result SVC model, while best reached 91.70% model.