Application of the GSB Process in Stochastic Modeling the Network Intrusion Detection Threshold
Vladica S. Stojanović, Mihailo R. Jovanović, Brankica M. Popovic, Petar Čisar, Kristijan V. Kuk · IPSI Transactions on Internet Research · 2025
This manuscript proposes one possible solution for intrusion detection (IDS) based on stochastic modeling of the threshold value, which is defined as a random variable (RV) obtained using the so-called General Split-BREAK (GSB) stochastic process. Applying such a model to previously recorded traffic values and using this type of stochastic modeling allows for more accurate and real-time threshold adjustment for IDS software. Conducted numerical simulations and application to real data show that alarms obtained by using this method are activated correctly and efficiently, with a potential reduction in the number of incorrect and fraudulent activations.