Sentiment Analysis of Users on Social Networks: Overcoming the challenge of the Loose Usages of the Algerian Dialect

Sentiment Analysis (SA) focuses on the study and analysis of peoples’ opinions, sentiments and emotions based on written language. It is currently a very active research area in NLP. The growth of Web 2.0 has given all internet users the additional power of “interactivity, interoperability, and collaboration [1]. This has rapidly opened the door for the development of social media and the large interaction between users in all walks of life. Social media platforms are currently exploited by many companies as a major channel to advertise and sell products. As such, tools are clearly needed to analyse peoples’ opinions on and reviews of the various products, feedback on events, etc. In recent years, researchers on Arabic NLP have made some good effort tackling the problem of SA. These efforts have been more focused during the last couple of years on Arabic dialects and, to a lesser extent, on the dialects of the Maghreb region, even less on the Algerian Dialect (AlgD). The processing of this dialect is made even more complex with the frequent code switching by its speakers between Arabic and Latin letters. Facebook being widely used in the Arab world, and in Algeria more specifically, we are interested in this paper in the SA of Algerian users’ comments on various Facebook pages. A painstaking pre-processing of a corpus of such comments is done, and two neural network models, MLP and CNN, are trained to classify comments as negative, neutral or positive. Though a complex dialect, we have obtained an 81.6% accuracy with the MLP network and 89.5% accuracy with the CNN. We find this as a very encouraging result.