ABSTRACTS OF ARTICLES OF THE JOURNAL "INFORMATION TECHNOLOGIES".
No. 6. Vol. 24. 2018

DOI:10.17587/it.24.406-413

D. V. Pantiukhin, Senior Lecturer, e-mail: dpantiukhin@hse.ru, E. Karelova, Student, e-mail: ekarelova@edu.hse.ru,
National Research University Higher School of Economics, Moscow

Quality Improvement of Intrusion Classification through Convolution Neural Network by Training on Balanced Samples

This research investigates the effects of training sample balancing while solving intrusion classification task with convolution neural network. Using two convolutional neural networks with similar architecture, we conduct comparative analysis of classification task solution quality with and without training sample balancing. Experiments illustrate the efficiency of using training sample balancing in case of significant differences in the amount of samples in different classes.
Keywords: intrusion detection, convolution neural network, training sample balancing, intrusion detection datasets

P. 406–413

To the contents