Detecting Smishing with NLP, Simulated User Behaviour, and Emotion Analysis: A Machine Learning Approach

O'Neill, Joshua (2025) Detecting Smishing with NLP, Simulated User Behaviour, and Emotion Analysis: A Machine Learning Approach. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (710kB) | Preview
Preview	PDF (Configuration Manual) Download (2MB) | Preview

Abstract

Smishing exploits urgency and emotion to trick users into clicking malicious links or disclosing credentials. Traditional filters that focus on URLs or blocklists struggle with short, rapidly evolving campaigns. This thesis presents an integrated smishing detection framework that combines contextual language modelling, emotion analysis and simulated user behaviour. Four public SMS corpora are merged into a unified dataset, pre-processed with consistent normalisation and evaluated under both balanced and realistically imbalanced splits. Classical TF-IDF baselines with Logistic Regression, Linear SVC and Multinomial Naive Bayes are compared against fine-tuned BERT and DistilBERT models. A late-fusion layer then concatenates GoEmotions probability vectors with BERT or TF-IDF smish scores and trains a Logistic Regression classifier, while a lightweight persona module translates outputs into risk ratings for different user types. Experiments show that BERT with a 256-token context achieves an F1 score of 0.9874 on a deployment-style test set, outperforming TF-IDF baselines. Emotion fusion delivers comparable accuracy while adding interpretable explanations, demonstrating that combining language, emotion and behaviour is a promising direction for human-centred smishing defence.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Mustafa, Raza Ul UNSPECIFIED
Subjects:	P Language and Literature > P Philology. Linguistics > Computational linguistics. Natural language processing Q Science > QA Mathematics > Computer software > Computer Security T Technology > T Technology (General) > Information Technology > Computer software > Computer Security Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Cyber Security
Depositing User:	Ciara O'Brien
Date Deposited:	04 Sep 2026 08:49
Last Modified:	04 Sep 2026 08:49
URI:	https://norma.ncirl.ie/id/eprint/9814

Actions (login required)

View Item