Detección de ataques realizados a un servidor por un nodo BOTNET mediante la implementación de un modelo de red neuronal de aprendizaje profundo (DEEP LEARNING) usando el conjunto de datos BETH

Abstract

This project presents the implementation of a deep learning neural network model for the detection of attacks on a server by a botnet node, using the BETH1 dataset. The context of the work is framed within the current problem of security attacks on networked computer systems and the need for intelligent control measures to protect them. The problem posed in this degree work is the detection of attacks from a botnet node to a server, since they cause significant damage to the network infrastructure, among them are: the generation of massive data traffic to a server, the sending of large unsolicited emails, the infection of network devices, among other damages. This problem is complex because botnet nodes are designed to evade intrusion detection systems and the traffic generated is similar to legitimate traffic. The developed solution consists in the implementation of a deep learning neural network model using the BETH dataset, which contains network traffic information generated by botnet nodes and legitimate traffic. The model performance is evaluated using performance metrics such as accuracy, recall and F1-score. The results obtained show that the proposed model achieves a botnet node detection accuracy of 91%, indicating that it is an efficient solution for detecting botnet attacks. In conclusion, the implementation of a deep learning neural network model for the detection of botnet node attacks on a server is an effective and efficient solution. The results obtained demonstrate the importance of using advanced data analysis and deep learning techniques to improve network computer security in the area of anomaly detection.