Cybercrime Detection through Min Max Ants System-Based Feature Selection and Classification Techniques

This research focuses on the comprehensive detection and prevention of cybercrimes, which encompass various criminal activities targeting computers and communication devices. These crimes range from child pornography and cyberbullying to identity theft, credit card fraud, hacking, and malware attacks. Such cybercrimes often lead to privacy breaches, security vulnerabilities, financial losses, money laundering, and damage to public and government assets. To address these challenges, the study explores various data mining techniques, including machine learning and deep learning, for cybercrime detection and prediction. Specifically, the research proposes the utilization of the Min-Max Ants System (MMAS) as a feature selection method, alongside popular machine learning techniques such as decision trees, random forests, support vector machines (SVM), boosting, linear regression, and neural networks. The study introduces a novel framework that combines MMAS feature selection with SVM, decision tree, random forest, boosting, and linear regression methods for classification. The experimental results demonstrate that the boost and SVM algorithms achieve the highest accuracy in cybercrime detection.