An Improved Flow Direction Optimization Algorithm for Spam Email Detection

With the advancement of science and technology, the popularity of the Internet, particularly email, has increased significantly. Email spam has become one of the most prevalent forms of cyberattack, primarily used to disseminate malicious content, including commercial advertisements, computer viruses, and misleading information. Cyber attackers often target systems and servers with various types of malware and viruses to compromise or gain unauthorized access to systems or email accounts. This paper presents an improved flow direction algorithm for feature selection and a k-nearest neighbor algorithm for email spam classification. The Flow Direction Algorithm (FDA) typically faces challenges such as getting stuck in local optima and lacking population diversity. To enhance the FDA's capabilities, chaos operators have been introduced to promote population diversity and expedite convergence. The proposed method employs two types of chaos: circular chaos and logistic mapping. The performance of the proposed model was evaluated using the Spambase dataset, which consists of 4601 samples and 57 features. The results demonstrate that the accuracy of the proposed model, particularly with logistic mapping, is higher than that of other methods.