Sentiment Analysis of Twitter Messages using Word2vec by Weighted Average

Terrorist groups and their supporters use social networks to incite terrorism. They spread their ideas and doctrines by sharing their opinions on these social networks. In this article, we propose a method for detecting radical content in Twitter. For this, we have collected Arabic tweets related to...

Full description

Saved in:
Bibliographic Details
Published in2019 Sixth International Conference on Social Networks Analysis, Management and Security (SNAMS) pp. 223 - 228
Main Authors Djaballah, Kamel Ahsene, Boukhalfa, Kamel, Boussaid, Omar
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.10.2019
Subjects
cross-validation
inciting terrorism
machine learning
prediction
Random forests
Sentiment analysis
social networks
Support vector machines
Terrorism
Twitter
Vocabulary
word2vec by weighted average
Online AccessGet full text
DOI10.1109/SNAMS.2019.8931827

Cover

More Information
Summary:Terrorist groups and their supporters use social networks to incite terrorism. They spread their ideas and doctrines by sharing their opinions on these social networks. In this article, we propose a method for detecting radical content in Twitter. For this, we have collected Arabic tweets related to terrorism activities, and classified some of them in two classes of sentiments, inciting terrorism or not inciting. We used Word2vec and Word2vec by weighted average to represent our tweets. In addition, we used two machine learning algorithms, namely SVM and Random Forest to predict sentiments. To validate our methods, we have done some experiments using cross-validation technique. To evaluate our results, we have used three measures. The results revealed that the use of Word2vec by weighted average is slightly better than Word2vec method.
DOI:10.1109/SNAMS.2019.8931827