Sentiment analysis for Arabic e-commerce websites
Mohamed Ali Sghaier, Mounir Zrigui · 2016
Sentiment analysis is a research area where studies focus on the analysis and the processing of opinions found on the web content. It is about finding the polarity of sentiment expressed by internet user in their interaction about a giving subject in order to classify them on positive or negative. In this paper, we propose the implementation of a tool for sentiment analysis able to find the polarity of opinions in reviews extracted from e-commerce magazines and blogs in Arabic language. To do this, we conducted various experiments like testing different techniques of stemming and the performance of some classification algorithms to have the combination that gives us satisfactory results. In spite of the huge difficulties that we found in this research area like the lack of resources in data collection and the complexity of processing the Arabic language dialects, the results were hopeful and satisfying.