Análise comparativa de algoritmos de inteligência artificial para detecção de smishing

• Autor(a) principal: Santos, Pedro Henrique dos
• Data de Publicação: 2024
• Outros Autores: Vasconcellos, Victor Huander Lima de
• Tipo de documento: Trabalho de conclusão de curso
• Idioma: por
• Título da fonte: Repositório Institucional PUCRS
• Texto Completo: https://hdl.handle.net/10923/26799
• topic: SMISHING; INTELIGÊNCIA ARTIFICIAL; APRENDIZADO DE MÁQUINA; DETECÇÃO DE PHISHING; ANÁLISE COMPARATIVA; CIBERSEGURANÇA; ARTIFICIAL INTELLIGENCE; MACHINE LEARNING; PHISHING DETECTION; COMPARATIVE ANALYSIS; CYBERSECURITY

Smishing, a deceptive form of cyber attack that combines SMS messages with phishing techniques, represents a growing threat to individuals and organizations. In this article, we compile a collection of necessary knowledge for a good understanding of the problem itself and some of its possible solutions. Additionally, we conduct a comparative analysis of two of the most modern Artificial Intelligence (AI) algorithms to evaluate their effectiveness in detecting smishing attempts in a diverse dataset of smishing messages. We explore popular machine learning models, including KNN, random forests, support vector machines, XGBoost, and LightGBM. Our research highlights the strengths and weaknesses of each algorithm in terms of accuracy, recall, and computational efficiency, providing valuable insights to enhance smishing detection systems.