Please use this identifier to cite or link to this item: http://hdl.handle.net/10609/97586
Title: Técnicas de aprendizaje automático para la detección de ataques en el tráfico de red
Author: Valencia Peral, Andrés
Director: Rifà-Pous, Helena  
Tutor: Hernández Jiménez, Enric
Abstract: The purpose of the project will be to carry out a general analysis of the currently available automatic learning techniques applied to the implementation of a network intrusion detection system. The different techniques existing today will be described in a simple way, highlighting the fundamental differences between classical Machine Learning techniques as opposed to those of Deep Learning using Neural Networks, as well as the underlying relationship between them. We will proceed with a more exhaustive study of one of the techniques of automatic learning, the decision tree, and finally we will deal with some detail of the aspects to be taken into account in the implementation of a neural network to tackle the problem, with special focus on the choice of training hyperparameters and the consequences that such decisions entail. We will conclude that due to the nature of the problem posed, which has extremely abundant sample sets to be used to train the desired models, the application of these techniques offers a decisive advantage over other traditional techniques based on rules and signatures.
Keywords: machine learning
deep learning
neural networks
IDS
Document type: info:eu-repo/semantics/masterThesis
Issue Date: 4-Jun-2019
Publication license: http://creativecommons.org/licenses/by-nc-nd/3.0/es/  
Appears in Collections:Bachelor thesis, research projects, etc.

Files in This Item:
File Description SizeFormat 
avalenciapTFM0619memoria.pdfMemoria del TFM3,41 MBAdobe PDFThumbnail
View/Open
Tester2.py7,63 kBPhoto CDView/Open