Strengthening Small and Medium-Sized Businesses’ Cybersecurity: A Machine Learning-based Phishing Classification Model
Open AccessSmall and medium-sized businesses (SMBs) are priority targets for cybercriminals due to lacking financial, human, and technical resources. Most SMBs cannot afford information security experts or cyber defense tools to defend against cyberattacks, necessitating cost-effective interventions to enhance cybersecurity. In 2022, cybercriminals caused over $10.3 billion in losses to SMBs using various cyberattack types like social engineering, phishing, and ransomware (FBI, 2023). Phishing is the top cyberattack for SMBs, often leading to more severe consequences like ransomware. A machine learning-based classification model can effectively determine phishing website attacks for SMBs. This praxis demonstrates that a cost-effective phishing detection tool can be developed using open-source tools, historical phishing data, and machine learning to detect phishing website attacks proactively. Using a novel ensemble feature selection technique, this praxis develops a features-based uniform resource locator (URL) parser to identify significant lexical features of phishing websites and evaluates which machine learning classifiers have the best performance for detecting phishing websites. The contributions of this praxis include the following: First, it demonstrates that the ensemble feature selection with the Random Forest classifier performed the best in classifying phishing websites with up to 97% accuracy. Second, it identifies and validates a global set of features using the latest phishing data sets. Finally, the research introduces a phishing data set for benchmarking machine learning models.
- All rights reserved
Notice to Authors
If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.