Jordan University of Science and Technology

Deep Recurrent neural network vs. support vector machine foraspect-based sentiment analysis of Arabic hotels? reviews

Authors:  Mohammad Al-Smadia, Omar Qawasmeh, Mahmoud Al-Ayyoub, Yaser Jararweh,Brij Gupta

tIn this research, state-of-the-art approaches based on supervised machine learning are presentedto address the challenges of aspect-based sentiment analysis (ABSA) of Arabic Hotels? reviews. Twoapproaches of deep recurrent neural network (RNN) and support vector machine (SVM) are implementedand trained along with lexical, word, syntactic, morphological, and semantic features. The proposedapproaches are evaluated using a reference dataset of Arabic Hotels? reviews. Evaluation results showthat the SVM approach outperforms the other deep RNN approach in the research investigated tasks (T1:aspect category identification, T2: aspect opinion target expression (OTE) extraction, and T3: aspect sentimentpolarity identification). Whereas, when focusing on the execution time required for training and testingthe models, the deep RNN execution time was faster, especially for the second task.