Mining Opinion and Sentiment from Arabic Text
Asif Malik, Samer Aoudi, Salem Alteneiji, Thair Khdour, Mohammed Saleh, Issam Hamdan · 2020
At a big data level there are voluminous amounts of posts in Arabic that can be found on social media and on various product/service review sites. This provides a rich source of information at a number of levels that can be useful both commercially and at government level. Analyzing media in English, which uses sentiment lexicons, is well established. Posts written in Arabic is a challenging task primarily due to Arabic’s rich morphology. There have been a number of efforts to build Arabic sentiment lexicons. However, they suffer from being of a limited size, unclear usability plan or publicly not available. This project, ASAM, aims to develop tools that will opinion mine media written in Gulf Arabic, Modern Standard Arabic and Arabizi (Arabish) or a mixture of these. In addition to mining opinion, the project will aim at demographically profiling the opinions.