Deep Learning-Based Sentiment Analysis of Algerian Dialect during Hirak 2019

Abstract

Recent studies have used sentiment classification to analyze social media content using either semantic rule-based methods such as linguistic extraction models obtained from morpho-syntactic analysis, statistical methods based on Bag-of-word techniques or machine learning and deep learning models. This work uses the traditional Machine Learning algorithms and the Deep Learning models (Convolutional Neural Networks CNN and Recurrent Neural Networks RNN) applying on corpus collected from social media (Facebook, YouTube and Twitter) about the Hirak_19 (popular protest in Algeria during 2019) written in Algerian Dialect to analyze sentiments and provide a deeper understanding of opinions. The corpus is built from several dialectal Arabic texts; it consists of 7800 comments about political Hirak topics. CNN and RNN have been applied in several applications; in this paper, we show their power of detecting and classifying opinions about a social issue and analyzing sentiments. The results are positive demonstrated by the accuracy scheme 63.28% (CNN) and 60.97 (RNN) of cross-validation tests.