Multi-Label Learning in Computer Networks
Oleg I. Sheluhin, Dmitriy I. Rakovskiy · 2023
An important problem of intelligent data processing of system logs is the existence of data sets containing records with multiple associations of class labels. Research objective: To formalize the problem of multi-label classification of experimental data (binary or multi-class) on the example of computer networks (CN) log records. Novelty: consists in illustrating the presence of multi-label class labels in the analysis of system log records generated by the CN. Results: It is shown that the problem of multivaluedness of system log class labels is relevant for the analysis of accessibility and integrity of information circulating in the CN.