![]() | Only 14 pages are availabe for public view |
Abstract The malicious and risky effects of the malware crisis on systems because of huge various types of malware and open scripts on the virtual world known as the World Wide Web. This work focuses on examining the most distinct types of malware and the methodologies of protecting ourselves from them by finding them and expelling them from the system. These methodologies are considered indirect defenders since these small pieces of software or code can be found everywhere within the client system. The first part of the thesis includes examining the literature review and data on malware harmful effects and the AI methods used in malware detection and classification through 2017-2023. An effective comparison is established based on algorithms, optimization criteria, real dispatch, malware types, and execution parameters. The second part of our work is new firewall frameworks. The proposed framework is composed of a system data-monitoring framework, a security framework against malware attacks, a security framework against database leak attacks, and a security framework against Trojan horses and worms based on an intelligent fake algorithm framework. Millions of malicious records are found each year, according to global reports. One of the most common reasons for such huge numbers of these records is that malware manufacturers regularly alter or mix insecure records from the same family, with the same malicious behavior, using a variety of techniques. These manufactured malware-distinct records are ranked to hardly be detected. We must be able to divide them into groups and distinguish their respective families based on their ordering behavior in order to viably analyze and classify these vast amounts of tissue. The globally documented Malimg malware dataset used in this reasoning with three new augmented classes of viruses using 131 tests. By training, validating, and testing the proposed system using CNN and SoftMax, We have achieved an accuracy of 99.7% compared with the highest global rank 98.48%. |