Desarrollo de un sistema inteligente de ciberseguridad basado en machine learning para la detección y mitigación de amenazas en redes empresariales simuladas.

Abstract

Abstract: This paper addresses the increasing complexity of cyber threats by developing a prototype intelligent system for threat detection in enterprise networks. Using a Generalized Linear Model (GLM) Machine Learning, the system will be designed to identify anomalous traffic patterns proactively. Several methodologies are applied in this work that allowed to manage the development by structuring the project in phases, ranging from understanding the problem to the implementation of the solution. A key feature is the inclusion of a simulation module that allows training and evaluation of the model in controlled environments, thus improving its ability to adapt to unknown attacks. Tests are also performed on the obtained datasets, thus demonstrating its high effectiveness and accuracy when performing classifications above 90%. The functional prototype integrates modules for data management, simulation, detection and visualization of alerts, thus validating an approach based on the integration of modules, in order to strengthen information security within companies.