The paper presents the method to forecast the states of elements of the Internet of Things based on using an artificial neural network. The offered architecture of the neural network is a combination of a multilayered perceptron and a probabilistic neural network. For this reason, it provides high efficiency of decision-making. Results of an experimental assessment of the offered neural network on the accuracy of forecasting the states of elements of the Internet of Things are discussed.

The paper discusses an approach for detecting cyber attacks against smart power supply networks, based on identifying anomalies in network traffic by assessing its self-similarity property. Methods for identifying long-term dependence in fractal Brownian motion and real network traffic of smart grid systems are considered. It is shown that the traffic of a telecommunication network is a self-similar structure, and its behavior is close to fractal Brownian motion. Fractal analysis and mathematical statistics are used as tools in the development of this approach. The issues of a software implementation of the proposed approach and the formation of a dataset containing network packets of smart grid systems are considered. The experimental results obtained using the generated dataset have demonstrated the existence of self-similarity in the network traffic of smart grid systems and confirmed the fair efficiency of the proposed approach. The proposed approach can be used to quickly detect the presence of anomalies in the traffic with the aim of further using other methods of cyber attack detection.

Analysis of computer network security is a serious challenge. Many security metrics has been proposed for this purpose, but their effective use for rapid and reliable security evaluation and generation of countermeasures in SIEM systems remains an important problem. The use of ontologies for security information representation in SIEM systems contributes largely to the success of this task. However, most of works on ontological security data representation does not take into account the ontologies of security metrics. This paper proposes a new approach on using security metrics which is based on their ontological representation and serves for comprehensive security evaluation and subsequent countermeasure generation. The novelty of the proposed approach is that ontology of security metrics is viewed as a core component of a countermeasure decision support system. The proposed solutions are tested on a specific example.

The paper proposes a new approach to solve role mining problem in role-based access control systems. This approach is founded on applying genetic algorithms as heuristic optimization methods that are effectively used when the search space is too huge to be fully explored. To realize genetic algorithms, we propose some important novelties: having many chromosomes by individuals, presentation of genes as complex objects, dividing selection and mutation into several phases, accounting data confidentiality and availability in fitness functions and other. Proposed genetic algorithms were tested on randomly generated data sets for "basic" and "edge" role mining problems. The test results allow to assert that genetic algorithms may be successfully applied to efficiently solve main kinds of role mining problems.

The paper presents an approach to detection of attacks against Internet-of-Things networks and devices which can be used in critical infrastructures. It is based on use of machine learning and big data processing. Feature of the offered approach is using the method of reduction of output data sets and application of various algorithms of machine learning based on distributed data processing. The paper compares the speed and accuracy of attack detection on the basis of machine learning algorithms in the local and distributed modes.

The paper proposes a traffic anomaly detection technique which could be implemented in networks of the Internet of things. It is based on using fuzzy logical inference applied to the stationary Poisson or self-similar traffic peculiar to networks of the Internet of things. The algorithms of the modified stochastic approximation and "sliding window", included in the traffic anomaly detection technique, are suggested. Results of an experimental assessment of the technique are discussed.