الفهرس | Only 14 pages are availabe for public view |
Abstract In recent years, due to the wide speed usage of the Internet, huge information cover peo- ple{u2019}s opinions has been produced over the web. People tend to post their views using Internet forums, discussion groups, product reviews, blogs and social media. Analyzing people{u2019}s opinions and extracting knowledge out of them are very important and challeng- ing. However, analyzing such huge information manually is time consuming and could be impossible. Opinion mining (OM) is a text mining discipline that automatically performs such a task. Most of the research done in this {uFB01}eld has focused on English texts with a very limited orientation to the Arabic language. This scarcity is due to speci{uFB01}c features that the Arabic language has. Generally, there are three levels of opinion classi{uFB01}cation in opinion mining, document, sentence and aspect. The sentence and document levels analysis do not discover what exactly people like or dislike. Aspect-based opinion mining solves the detailed information problem. Extracted aspects and their polarity provide people with correct information and help them to take right decisions. However, studying the opinion text, especially on aspect level, is extremely challenging |