Title: Arabic language sentiment analysis via cross-language translation

Authors: Fahad Kamal Alsheref

Addresses: Information Systems Department, Beni-Suef University, Egypt

Abstract: Social platforms on the web give us a huge amount of data like comments, messages, products' reviews and posts. Sentiment analysis on social users' data is a valuable analysis for automatically understanding people's opinions about products or some interesting topics, and this analysis provides valuable information for informed decision making in different domains. Arab web users number is about 200 million users; this huge number represents the strong impact on marketing which makes the sentiment analysis of Arabic language hot topic. This paper proposed an Arabic sentiment analysis model to increase the accuracy of Arabic sentiment analysis. It is based on a cross-language retrieval model. The proposed model was tested on a dataset extracted from twitter and manually sentimentally classified by experts, and it achieved 90.62% accuracy.

Keywords: the social network; cross-language; information retrieval correlation coefficient; Facebook; Twitter; sentiment analysis; social network service; text mining.

DOI: 10.1504/IJWBC.2019.101047

International Journal of Web Based Communities, 2019 Vol.15 No.2, pp.151 - 159

Received: 30 Aug 2018
Accepted: 01 Jan 2019

Published online: 22 Jul 2019 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article