Publication Date: 2024/01/09
Abstract: The internet has grown to be a significant part of our daily lives for communication and dissemination of information. It facilitates the unexpected and seamless execution of knowledge. Understanding break-ins and individual deception as different parts of bad behavior mean that software engineers and bad customers get the personal information of present good customers to try deception or foxy motivation for unauthorized related gain. Malignant URLs have unconstrained substance (like junk mail, phishing, pressure-by-using exploits, and so forth) and phishing the customer to turn out to be losses from stunts (financial adversity, theft of personal facts, and malware basis), furthermore, cause mishaps of bil- lions of bucks reliably. To resolve the existing problem, we have proposed the algorithm Linear Support Vector Machine (LSVM) with One-against-approach and Convo- lutional Neural Network (CNN). In this study, we use a clear estimation to determine whether a URL is good or bad. When compared to the existing study, our proposed method attained more accuracy.
Keywords: Malicious Detection, Cyber Security, Big Data, CNN, Linear SVM, URL Dataset, Classification.
DOI: https://doi.org/10.5281/zenodo.10474017
PDF: https://ijirst.demo4.arinfotech.co/assets/upload/files/IJISRT23DEC1796.pdf
REFERENCES