Search academic texts

Information × Registration Number 0224U002653, 0121U112950 , R & D reports Title Fulfillment of the tasks of the perspective plan of development of the scientific direction "Military sciences and national security" popup.stage_title Head Popkov Borys O., Кандидат військових наук Registration Date 22-02-2024 Organization Taras Shevchenko National University of Kyiv popup.description2 The object of the study is the process of functioning of the information system under the influence of cyber attacks. The subject of research is the methods of intelligent distribution of Data Mining data. The task is to create an intrusion detection system based on the use of Data Mining methods. A formalization of the intrusion detection system model has been developed. This model is based on a multitude of diverse methods of intelligent data distribution. The presented model of the system enables the detection of the majority of network attacks and anomalies by leveraging broad capabilities for forming detection modules. A methodology for conducting experimental research on the application of Data Mining methods in the tasks of detecting network attacks has been developed and practically implemented. As a result, an experimental study was conducted, confirming the correctness of the proposed model and the selection of numerous intelligent data distribution methods underlying it. The Support Vector Machines method allowed the identification of the majority of attacks with a success rate of 98–100%. The Principal Component Analysis method reduced the amount of information required for classifying network packets and increased the speed of forming detection modules but revealed an issue of overfitting. Clustering methods allowed the formation of numerous detection modules, isolating typical attack fragments into separate detection modules and breaking down complex attacks into individual modules. The application of fuzzy logic improved the system's performance and enabled the classification of vectors with different labels in the training sample. In summary, the conducted experimental research confirmed the accuracy of the proposed model and the selection of various intelligent data distribution methods underlying it. Product Description popup.authors Toliupa Serhii V. popup.nrat_date 2024-02-22 Close