Feature-Based Sentiment Analysis in Online Arabic Reviews

Social media has given web users a venue for expressing and sharing their thoughts and opinions on different topics and events. Each day millions of user generated comments are raised on the web and analyzing these opinions to discover useful information pieces manually is costly and time-consuming. Thus, automatic mining techniques are highly desirable. By investigating the current state of automatic sentiment analysis tools, a lack of tools for analyzing languages rather than English was highly observed. Most of the researches on Opinion Mining are tailored for English language, and research on mining Arabic reviews is going in very slow rate. This study proposes a feature-based sentiment analysis technique for mining Arabic user generated reviews. The extraction and weighting of sentiments and features are executed automatically from a set of annotated reviews using Part Of Speech (POS) tagging feature. The collected features are organized into a tree structure representing the relationship between the objects being reviewed and their components. Furthermore, an automatic expandable approach of Arabic feature and sentiment words using free online Arabic lexicons and thesauruses is introduced. For extracting and analyzing feature-sentiment pairs five rules is proposed. Finally, a lexicon-based classification is performed to evaluate the performance of each rule. The experimental results show that the proposed approach is able to automatically extract and identify the polarity for a large number of feature-sentiment expressions and achieve high accuracy.