Arabic Sentiment Analysis Based Machine Learning for Measuring User Satisfaction with Banking Services' Mobile Applications : Comparative Study

Understanding the quality of the application and the user's needs are important tasks in application development. To understand user requirements to increase application quality, application review-based sentiment analysis (SA) can be used. This research aimed at determine customer opinions toward mobile banking services' applications, then update and maintain these applications. Banking services' mobile apps became a significant component of people life. So, these mobile apps user comments have targeted in this study for SA tasks. Utilized dataset is collected from banking mobile apps user reviews on Google Play Store. Labeling process is performed manually, it generate three main classes: positive, negative and neutral classes. Arabic sentiment analysis process is achieved using machine learning techniques, namely Naïve Bayes (NB), K-nearest neighbor (KNN), Decision Tree (DT) and Support Vector Machine (SVM) models. The NB model has gained a better level of evaluation comparing to other DT, KNN and SVM algorithms. NB model has achieved the highest accuracy, recall, precision and F-score, which were 89.65% in accuracy, 88.08% in recall, 88.25% in precision, and 88.25% in f-score evaluation measure.