CHOICE OF FEATURE SPACE FOR CLASSIFICATION OF NETWORK IP-TRAFFIC BY MACHINE LEARNING METHODS

IP-protocol and transport layer protocols (TCP, UDP) have many different parameters and characteristics, which can be obtained both directly from packet headers and statistical observations of the flows. To solve the problem of classification of network traffc by methods of machine learning, it is necessary to determine a set of data (attributes), which it is reasonable to use for solving the classification problem.

Hali tarjima qilinmagan