Comparison of Classification Algorithm Accuracy Levels in Phishing Detection

ABSTRACT Phishing attacks have become a serious threat in the realm of cybersecurity, with both direct and indirect impacts on individuals and organizations. This study aims to assess the performance of three machine learning classification algorithms—Decision Tree, Random Forest, and Support Vector Machine—in detecting phishing attempts through the use of suitable datasets. The study focuses on comparing the performance of these algorithms to determine their accuracy and reliability in phishing detection. The experimental findings indicated that Random Forest attained the highest accuracy, reaching 96.675%, followed by Decision Tree (93.57%), while Support Vector Machine showed lower performance (54.724%). Cohen's Kappa also indicated that Random Forest had the highest correlation value. In conclusion, for this dataset, Random Forest and Decision Tree are more effective at detecting phishing attacks compared to Support Vector Machine.