Opinion Mining for Arabic Customer Feedback Using Machine Learning

Mohamed M. Abbassy and Ayman Abo-Alnadr

The aim of opinion mining analysis is to classify the viewpoints of users through text feedbacks. Nonetheless, it is a difficult task to extract these views from huge quantities of unstructured and lengthy documents. Such types of information tools are known as important intelligence suppliers for big companies, economies, news, and many others and are an expanded forum for human emotions, thoughts, feedbacks, and judgments. Including the growing number of customers provide products in Arabic countries, the great value of these businesses involves understanding Arabic feedback across e-commerce applications and websites. In this paper, has been recommended techniques that benefit from an e-commerce website and application analysis valuable opinions and views to develop opinion mining in Arabic. Four techniques for classification such as Vector Support Machine (SVM) and Back Propagation Neural Networks (BPNN), Naïve Bayes, and Decision Tree, are introduced in this paper. The main purpose of this paper is to find an easy-to-use method for Arabic-language feedback from e-commerce websites. Results show that the maximum precision rating for SVM classification is 96.06 percent relative to other classificatory.

Volume 12 | 03-Special Issue

Pages: 209-217

DOI: 10.5373/JARDCS/V12SP3/20201255