• JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
 
  Bookmark and Share
 
 
Master's Dissertation
DOI
https://doi.org/10.11606/D.55.2023.tde-12012024-154719
Document
Author
Full name
Nicole do Vale Dalarmelina
E-mail
Institute/School/College
Knowledge Area
Date of Defense
Published
São Carlos, 2023
Supervisor
Committee
Meneguette, Rodolfo Ipolito (President)
Affonso, Frank José
Gonçalves, Vinícius Pereira
Nakamura, Luis Hideo Vasconcelos
Title in Portuguese
Uma abordagem Ensemble Learning para modelos de detecção de intrusão para redes industriais
Keywords in Portuguese
Aprendizado de máquina
Aprendizado em conjunto
IDS
IIoT
Abstract in Portuguese
A Internet tem se tornado um recurso essencial para a humanidade e para os dispositivos tecnológicos existentes atualmente, tanto dentro de casa Internet of Things (IoT) quanto dentro de industrias Industrial Internet of Things (IIoT). Todo esse avanço tecnológico pode trazer benefícios, mas também pode oferecer riscos à integridade dos dados se a segurança não for devidamente realizada utilizando Sistemas de Detecção de Intrusão (IDS) eficientes. Neste trabalho é proposto um modelo que poderá ser utilizado por IDSs para redes industriais utilizando Ensemble Learning. Para isso são analisadas abordagens para a extração das melhores features dos datasets utilizados, assim como a aplicação de algoritmos de balanceamento de dados a fim de selecionar as melhores abordagens para o treinamento do modelo proposto viabilizando possíveis retreinamentos do modelo a cada novo ataque encontrado, o modelo desenvolvido no presente trabalho obteve acurácia de 99.93%, concluindo seu treinamento em apenas 1 hora e 34 minutos, enquanto o modelo treinado utilizando os datasets sem tratamento obteve acurácia de 99.94% concluindo seu treinamento em 156 horas.
Title in English
An Ensemble Learning approach for intrusion detection models for industrial networks
Keywords in English
Ensemble Learning
IDS
IIoT
Machine Learning
Abstract in English
The Internet has been turned into an essential resource to humanity and to currently existing technological devices, both indoors Internet of Things (IoT) and in industrial environments Industrial Internet of Things (IIoT). All this progress can bring benefits, yet it can also bring risks to the data integrity if the security has not been properly performed by using effective Intrusion Detection Systems (IDS). In this work it is proposed the training of a model that can be used by IDS industrial networks using Ensemble Learning. For this intent, approaches for extract the best features of the used datasets are analyzed, as well as the use of data balancing algorithms in order to select the best approaches for training the proposed model, enabling possible retraining of the model for each new found attack, the model developed in the present work obtained an accuracy of 99.93%, completing its training in just 1 hour and 34 minutes, while the model trained using the datasets without treatment obtained an accuracy of 99.94%, concluding its training in 156 hours.
 
WARNING - Viewing this document is conditioned on your acceptance of the following terms of use:
This document is only for private use for research and teaching activities. Reproduction for commercial use is forbidden. This rights cover the whole data about this document as well as its contents. Any uses or copies of this document in whole or in part must include the author's name.
Publishing Date
2024-01-12
 
WARNING: Learn what derived works are clicking here.
All rights of the thesis/dissertation are from the authors
CeTI-SC/STI
Digital Library of Theses and Dissertations of USP. Copyright © 2001-2024. All rights reserved.