DETECT THE TENTATIVE BEFORE BECOMING REAL: A MACHINE LEARNING APPROACH FOR PHISHING EMAIL DETECTION IN ROMANIAN HEALTHCARE
DETECT THE TENTATIVE BEFORE BECOMING REAL: A MACHINE LEARNING APPROACH FOR PHISHING EMAIL DETECTION IN ROMANIAN HEALTHCARE

Author(s): George B. Mertoiu, Gabriela Mesnita
Subject(s): Economy, Business Economy / Management, ICT Information and Communications Technologies, Socio-Economic Research
Published by: Editura Universităţii »Alexandru Ioan Cuza« din Iaşi
Keywords: phishing detection; machine learning; healthcare; Romania;
Summary/Abstract: Phishing attacks pose a significant threat to individuals and organizations, and their accurate and effective detection is crucial to preventing data breaches and financial losses. With the increasing use of email as a communication channel, phishing attacks have become more widespread and sophisticated. Our study addresses the use of machine learning-based models to detect phishing emails by analyzing the text of the message. A characteristic of the study is given by the fact that it uses a dataset composed of private emails in Romanian, obtained from public institutions in the field of health. Since the models were applied to the text, natural language processing techniques specific to the Romanian language were used to extract the features. The results obtained highlighted that some models outperform others in terms of accuracy, underlining the importance of choosing a machine learning approach for phishing detection in a given language. The conclusions of this study can support research for the development of effective phishing detection tools for Romanian-speaking users and organizations.

• Details
• Contents