Dark Web Text Classification by Learning through SVM Optimization

The Darkweb has become the largest repository of unauthorized information compared to the surface web because of its benefit of anonymity and privacy. With these anonymity and privacy features, the dark web is also becoming a safe place for illegal activities and hence an increase of dark web usage and size of the onion-based URLs. With the increasing use of dark web users, it is the need for cybercrime investigators across the globe to classify dark web data for understanding various illegal activities to control and categorize URLs hosting such illicit activities with feature engineering. In this research, the Support Vector Machines (SVM) algorithm is used to understand the algorithm's efficiency for a proposed model to classify dark web data with optimization techniques. Text-based keywords from more than 1800 websites were collected by applying feature engineering techniques and the system's performance was evaluated with the SVM approach. The results are very encouraging as the Precision, Recall, and F-measure values are 0.83, 0.90 & 0.96 achieved with a dataset of 1800 URLs.