Table of Contents Author Guidelines Submit a Manuscript
The Scientific World Journal
Volume 2014 (2014), Article ID 631394, 8 pages
http://dx.doi.org/10.1155/2014/631394
Research Article

A System for Sentiment Analysis of Colloquial Arabic Using Human Computation

Information Technology Department, College of Computer and Information Sciences, King Saud University, Riyadh 11496, Saudi Arabia

Received 27 December 2013; Accepted 2 April 2014; Published 29 April 2014

Academic Editors: G. Wei and F. Yu

Copyright © 2014 Afnan S. Al-Subaihin and Hend S. Al-Khalifa. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

We present the implementation and evaluation of a sentiment analysis system that is conducted over Arabic text with evaluative content. Our system is broken into two different components. The first component is a game that enables users to annotate large corpuses of text in a fun manner. The game produces necessary linguistic resources that will be used by the second component which is the sentimental analyzer. Two different algorithms have been designed to employ these linguistic resources to analyze text and classify it according to its sentimental polarity. The first approach is using sentimental tag patterns, which reached a precision level of 56.14%. The second approach is the sentimental majority approach which relies on calculating the number of negative and positive phrases in the sentence and classifying the sentence according to the dominant polarity. The results after evaluating the system for the first sentimental majority approach yielded the highest accuracy level reached by our system which is 60.5% while the second variation scored an accuracy of 60.32%.