Classification and Identification of Phishing Websites based on Machine Learning

Phishing is the largest network security issue among global cybercrimes in 2022. Its frequency of occurrence has maintained rapid growth and has become one of the most important network security issues. In the state of the art of this research field, there was a trade-off between high-precision discriminant models and huge consumption of computing resources. Therefore, the research purpose of this article is mainly to balance the relationship between accuracy and computing resources (performance) to achieve accuracy and computing efficiency at the same time. This article uses principal component analysis (PCA) as a tool, uses its excellent dimensionality reduction ability to process sample data, compresses the original feature set, and then uses different machine learning models to conduct experiments. In the end, the random forest model after PCA achieved a discrimination accuracy of 97.157% with a performance improvement of 25.1%, effectively achieving a win-win balance between accuracy and performance.