A genetic programming approach to feature selection and construction for ransomware, phishing and spam detection
Harith Al-Sahaf, Ian S. Welch · Proceedings of the Genetic and Evolutionary Computation Conference Companion · 2019
Feature selection and construction can potentially help reduce dimensionality and build more effective features that aim to improve performance. This paper utilises Genetic Programming (GP) to automatically select and construct high-level features for three cybersecurity-related tasks namely Ransomware detection, Spam detection, and Phishing website detection. The effectiveness of the features constructed by the proposed method has been assessed using three commonly-used machine learning algorithms on three datasets and compared against the performance of these machine learning algorithms applied to the original set of features and those features selected and constructed by another GP-based. The experimental results show that the proposed method has significantly improved the performance compared to the other methods.