Mostra i principali dati dell'item

dc.contributor.authorFarsimadan, Eslam
dc.date.accessioned2024-07-15T08:18:49Z
dc.date.available2024-07-15T08:18:49Z
dc.date.issued2023-06-26
dc.identifier.urihttp://elea.unisa.it/xmlui/handle/10556/7257
dc.description2021 - 2022it_IT
dc.description.abstractThe Internet and advanced communication networks, such as IoT and cellular net- works, produce enormous and diverse traffic data flows. The behavior of network traffic in these networks is highly intricate due to factors like device mobility and network heterogeneity. As a result, conventional network security and management methods struggle to handle the challenges of securing, monitoring, and analyzing the network and data. These challenges include issues such as the efficacy of classi- fication and detection strategies, precision, accuracy, and the ability to process big data in real-time. Recently, machine learning and deep learning have proven to be highly effective in addressing network security concerns and have demonstrated superiority over tradi- tional methods. Consequently, researchers in the field of networking are turning to these machine learning and deep learning models for network security and manage- ment. Motivated by the success and effectiveness of these models, this thesis concentrates on addressing two crucial and challenging issues in network security through dynamic analysis, machine learning, and deep learning models, with a particular emphasis on artificial neural networks. More precisely, it presents new learning-based techniques for attack classification and malware detection. First, a general introduction with some motivations for this thesis is presented. Then, the state-of-the-art is investigated, and the most effective artificial intelligence- based methods related to the above mentioned network security aspects are reported. After that, the primary materials and preliminaries for constructing the proposed models, such as neural network types, recurrence plots, and so on, are introduced in detail. Finally, the proposed methods for attack classification and malware de- tection are investigated in terms of mathematical background, model architecture, experimental setting, and evaluation. Moreover, the proposed methods are analyzed from a theoretical perspective and through specific performance evaluation experiments on real network traffic datasets. The obtained results, in the presence of several unbalanced datasets instances, prove the effectiveness of the proposed approaches. Keywords: Network Security, Attack Detection, Attack Classification, Mal- ware Detection, Machine Learning, Deep Learning, Neural Networks, Recurrence Plots, Markov Chain. [edited by Author]it_IT
dc.language.isoenit_IT
dc.publisherUniversita degli studi di Salernoit_IT
dc.subjectNetwork securityit_IT
dc.subjectAttack detectionit_IT
dc.subjectAttack classificationit_IT
dc.subjectMalware detectionit_IT
dc.titleA Study of some ML and DL-based Strategies for Network Securityit_IT
dc.typeDoctoral Thesisit_IT
dc.subject.miurINF/01 INFORMATICAit_IT
dc.contributor.coordinatoreDe Lucia, Andreait_IT
dc.description.cicloXXXVit_IT
dc.contributor.tutorPALMIERI, Francescoit_IT
dc.identifier.DipartimentoInformaticait_IT
 Find Full text

Files in questo item

Thumbnail
Thumbnail

Questo item appare nelle seguenti collezioni

Mostra i principali dati dell'item