PREDICTION OF THE PROFILE FUNCTIONING OF A COMPUTER SYSTEM BASED ON MULTIVALUED PATTERNS

Abstract

Purpose of work is to create a new algorithm for predicting anomalous states of computer systems (CS) using the mathematical apparatus of multivalued dependencies (Multivalued Dependencies Prognosus Algorithm, MDPA), which are categorical concepts. The research method is the analysis of historical data using the mathematical apparatus of multivalued dependencies. Objects of study are theoretical and practical issues of solving and visualizing information security problems. Results of the study. A methodology and algorithm for predicting the state of CS have been developed. The boundaries of the input parameters of the algorithm are derived and justified. The boundaries of the input parameters need to be pre-configured for the correct generation of the prognosis. A software implementation of the proposed prediction algorithm has been developed. The efficiency of the algorithm has been tested on real experimental data. A spatial analysis of the prediction results was carried out. The main disadvantage of the proposed algorithm is the need to fine-tune the input parameters for each set of “historical data”. Scientific significance. The scope of application of multivalued dependencies has been expanded; a new algorithm for predicting anomalous states of CS, which are categorical concepts, has been proposed. The developed prediction algorithm can be generalized to any subject area containing historical data of any type