Model Arabic Sentiment Analysis Using Different Machine learning Algorithms in Hotels Domains

Abstract

With the popularity of Social Medias, sentiment analysis turns into the most important technique for understanding the opinion of users toward products or services. Arabic users are writing their comments using unstructured non-grammatical colloquial Arabic language, which made a complex challenge for sentiment analysis to operate without doing a lot of cleaning preprocessing stage. In this respect, this paper proposes a model on Arabic sentiment analysis with different classification algorithms using manual Arabic lexicon in hotels domain. Our experiment shows Logistic Regression and Support Vector Machine has the highest accuracy with 89% than other classifiers.