Investigating Deep Learning Word2vec Model for Sentiment Analysis in Arabic and English languages for User’s reviews

Abstract

In this paper, we explore natural language processing (NLP) methods to perform sentiment analysis or opinion mining. In addition, we show an application on English and Arabic sentiment analysis by implementing sentiment classification for three datasets which are Booking hotel dataset, Food fine Amazon dataset and Arabic movie review dataset. We applied Word2Vec model followed by Random Forest classifier (RF) for Arabic movie dataset. The results show that the Word2Vec model shows highly effective performance in sentiment analysis for English language datasets but it does not work for Arabic language as Arabic language need different mechanism.