Please use this identifier to cite or link to this item: http://dspace.univ-guelma.dz/jspui/handle/123456789/13258
Full metadata record
DC FieldValueLanguage
dc.contributor.authorBRAHMIA, ABDELBACET-
dc.date.accessioned2022-10-13T14:01:37Z-
dc.date.available2022-10-13T14:01:37Z-
dc.date.issued2022-
dc.identifier.urihttp://dspace.univ-guelma.dz/jspui/handle/123456789/13258-
dc.description.abstractIntrusion detection systems (IDS) are the subject of many studies and play an important role in network security. The purpose of this study is to model such a system to help system administrators detect and identify security breaches in their organization so that they can be prevented before they cause harm or damage. For this, we studied the performance of the learning methods machine (ML) applied to intrusion detection for cybersecurity. Then, we applied two detection techniques based on deep learning approaches, a deep neural network (DNN), and a convolutional neural network (CNN) to detect intrusions into network connections. We evaluated the proposed methods with the Edge_IIoT dataset of realistic cyber security traffic of IoT and IIoT attacks on networks. We also have presented realistic machine learning projects, we calculate and evaluate our work using different metrics applied for performance evaluation machine and deep learning (Precision, Recall, F1 score), and other important performance indicators for intrusion detection (confusion matrix ). The experimental results showed that the performances of the approaches of deep learning (DL) is depend on the database that you use and on the model you choseen_US
dc.language.isofren_US
dc.publisherUniversité de Guelmaen_US
dc.subjectCybersécurité, Système de détection d'intrusion (IDS), Deep Learning, Machine Learning , Edge_IIoTen_US
dc.titlel'apprentissage en profond pour la cybersécurité dans l'Internet industriel des objets (IIoT)en_US
dc.typeWorking Paperen_US
Appears in Collections:Master

Files in This Item:
File Description SizeFormat 
BRAHMIA_ABDELBACET_F5.pdf1,96 MBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.