Effectiveness of Social Media analytics counterterrorism Technology (smact) on Terrorism containment rates in Lamu county beach tourism destination, Kenya
Date
2024
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Chuka University
Abstract
Fear of terrorism has led to social and economic decline particularly in beach tourism destinations. The main
approaches used in managing terrorism include hard and soft strategies such as military and counter-radicalization.
Previous studies have reported the capabilities of social media analytics in detecting terror threats through media
streams. However, the effectiveness of this technology has not been examined in beach tourism destinations. This study
examines the effectiveness of social media analytics counterterrorism technology (SMACT) in identifying terror threats
in tourism destinations. The aim of the research is to determine the effectiveness of SMACT technology on terrorism
containment rates in Lamu County beach tourism destination. Descriptive research and machine learning is used to
process Twitter dataset using Python. A dataset of 9,572 tweets is preprocessed. Naive Bayes and Recurrent Neural
Network (RNN) sentiment analysis models are implemented and evaluated to categorize terrorism-related tweets as
positive or negative. A dataset of 5840 tweets is processed after preprocessing and split 80:20 into train and test sets
models are developed and trained. The trained models are used to predict sentiment on the 20% test set. A total of
1168 tweets counts are predicted on the test set for each model. Naive Bayes model predicts 719 positive and 449
negative tweets. The RNN model predicted 829 positive and 339 negative tweets. The Naive Bayes and RNN models
demonstrate highly accurate detection of extremist sentiment in tweets, though the Naive Bayes classifier outperformed
the RNN model. The Naive Bayes model predicted 61.6% of tweets as positive sentiment and 38.4% as negative. In
contrast, the RNN model predicted a 71:29 split between positive and negative tweets. Frequently occurring terrorism
related terms include 'attack', 'kill', 'bomb' among others. Twitter terrorism and counterterrorism tweets sentiments are
successfully analyzed, giving important insights into the prevalent opinions on the platform. The sentiment distribution
and correlations between positive and negative feelings are clustered. The results help researchers, policy makers and
other stakeholders in counterterrorism efforts gain a better grasp of public opinion on Twitter.
Description
Article
Keywords
sentiment analysis, Twitter, terrorism, counterterrorism, social media, dataset, and data preprocessing.
Citation
K. Biwott G , M.Muruiki L. and C. Too . (2024). Effectiveness of Social Media analytics counterterrorism Technology (smact) on Terrorism containment rates in Lamu county beach tourism destination, Kenya In: Mutembei Henry, Nduru Gilbert, Munyiri Shelmith, Gathungu Geofrey, Kiboro Christopher, Otiso Wycliffe, Rithaa Jafford, Miriti Gilbert, Gichumbi Joel, Mwathi David, Gitonga Lucy, Nanua Jackin, Kahindi Roseline, Jonathan Kathenge & Muthui Zipporah (Eds.). Proceedings of the Chuka University Tenth Annual International Research Conference held in Chuka University, Chuka, Kenya from 5th to 6th October, 2023.314-321 pp.