| Abstract [eng] |
The final master thesis examines intrusion detection systems and newest intrusion detection methods. It also contains explanation of machine learning and various types of learning algorithms. In addition, data collection methods are reviewed along with general steps for data collection. Furthermore there is a short review of both commercial and non-commercial intrusion detection systems. A method for intrusion detection is proposed which is based on a type of deep neural network. A data collection tool was programmed specifically for the chosen method and data was collected in an unsimulated environment. Last of all a prototype of selected method was programmed and it's effectiveness was evaluated by comparing it to other deep learning methods and data. Structure: introduction, analytical part, suggested intrusion detection method with data collection and evaluation of results, conclusion. Thesis consists of 61 pages of text, 40 pictures, 5 tables, 30 sources. |
