Sentiment Analysis of Arabic Algerian Dialect Using a Supervised Method

Sentiment analysis holds an important place in Natural Language Processing (NLP) due to its utility in resolving different issues in many fields such as e-commerce, politic sciences, social media analysis, cybersecurity, etc. In fact, most of the work done in the field has been dedicated to the sentiment of English texts. However, adapting research findings to the Arabic language and its dialects is not trivial because of their linguistic features. In this work, we apply two supervised methods, namely, Deep Learning and Support Vector Machines (SVM), for sentiment analysis of Modern Arabic and the Algerian dialect. These methods have been applied on a huge annotated dataset collected from different Arabic Algerian sources. Findings are showing promising results in sentiment analysis of the Arabic Algerian dialect.