Proposed model to balance between accuracy and efficiency in the detection of phishing: An approach that combines clustering and random forest

Abstract

1. Abstract The purpose of this study is to improve the identification of phishing attempts by us-ing a complete method that integrates clustering pre-pro-cessing with efficient Random Forest training techniques. Applying clustering algorithms to a well selected phishing dataset enables the identifi-cation of patterns and the re-finement of features, resulting in enhanced accuracy in de-tecting phishing attempts. The research concurrently investi-gates methods to decrease the training duration of Random Forest, such as modifying the quantity of weak learners, us-ing sampling approaches, and examining different fusing pro-cedures. The study endeavors
to achieve a harmonious equi-librium between precision and effectiveness via a process of repeated refinement. The re-sults enhance the area of cy-bersecurity by providing valu-able insights into the effective use of clustering and Random Forest training to mitigate phishing threats with greater resilience.