Max Perera, Jorge Carlos2015-08-172015-08-1701/07/2004http://hdl.handle.net/11285/572110The area of Intrusion Detection is very important these days. Companies have acquired more interest in having this type of systems beacuse of the importance that information has for them. Machine learning algorithms are being used along with IDSs as an efficient approach. For these reasons we work with this approach in this thesis, presenting from general to specific, the information of the models and types of IDSs, and some machine learning algorithms and some fusion rules for them, that can help achieving a good IDS. In this work, we focus on Host-based intrusion detection, and three machine learning algorithms, which are C4.5, RIPPER and PART. It is showed a method to reduce false alarm rates and with this, increasing the possibility of detecting true alarms when our system trigger them.info:eu-repo/semantics/openAccesshttp://creativecommons.org/licenses/by-nc-nd/4.0Area::INGENIERÍA Y TECNOLOGÍA::CIENCIAS TECNOLÓGICAS::TECNOLOGÍA DE LOS ORDENADORES::LENGUAJES ALGORÍTMICOSTesis de maestríaIDSHost-based IDSLow Overhead IDSTelecommunicationsElectronic EngineeringIngeniería y Ciencias Aplicadas / Engineering & Applied Sciences