This paper introduces a supervised ap- proach for performing sentence level di- alect identification between Modern Stan- dard Arabic and Egyptian Dialectal Ara- bic. We use token level labels to de- rive sentence-level features. These fea- tures are then used with other core and meta features to train a generative clas- sifier that predicts the correct label for each sentence in the given input text. The system achieves an accuracy of 85.5% on an Arabic online-commentary dataset outperforming a previously proposed ap- proach achieving 80.9% and reflecting a significant gain over a majority baseline of 51.9% and two strong baseline systems of 78.5% and 80.4%, respectively.

Click Here to View the Paper In XML Format