Analysis and classification of Arabic crowd-sourced news reports: A case study of the Syrian crisis

The prevalence of social media, in the whole world and the Arab region in particular, has fueled the active engagement and participation of large swaths of the Arabic society in current events. Social media has been used to rally public opinion, increase awareness, spread information/misinformation, and organize large events. Data analysis is necessary to drive decision making, advertisement, political campaigning, counter-intelligence, etc. The sheer volume of data and number of users calls for automated methods for analysis and classification of Arabic text. In this paper, the problem of analysis and classification of Arabic news reports was studied. Innovative methods, based on lexical analysis and machine learning, were employed to tame the complexity of the Arabic language. Different classification algorithms were compared and the classification accuracy results are promising. This research presents seminal steps toward specialized analysis of Arabic crowd-sourced data and social media.