Twitter Sentiment Analysis and Events Prediction Using Lexicon Sentiment Analysis for Iraqi Vernacular

Authors

  • Ali A. Tuama, Ahmed T. Sadiq

Abstract

Recently, using of social media platforms are ranked high in terms of people communication, opinions exchanging, publishing and circulating news and events; so, it is become necessary to increase the scientific means to search for the contents of these platforms to obtain information which is useful in several areas like trend analysis, determining people’s criteria, pages, posts, comments, and tweets.

In this research, a sentiment analysis system was proposed that would benefit decision makers in supporting their decisions regarding positive, negative, and neutral tendencies. The classical Arabic language was adopted as well as the colloquial Iraqi dialect due to its importance in the frequent circulation of social networking pages and sites. A complete database has been built for the important words that play a role in distinguishing the positive and negative tendencies of the Iraqi dialect with the creation of database of people information from Twitter. Three types of important classification methods were used for the purpose of categorizing tendencies (Rough Set Theory, Naïve Bayes, K-NN). The main purpose is to categorize tendencies and the other purpose is to make a comparison between the three methods. Where the results show that the (Rough Set Theory) method is the best classification method for tendencies with a success rate of 97% on the Tweeter social network.

Published

2020-11-20

Issue

Section

Articles