Results of Using Fuzzy Neural Subnetworks Method in Intelligent Data Analysis of Internet of Things Image Sensors

The article presents the results of an analytical study on the use of the fuzzy neural subnetwork method in the intelligent analysis of IoT image sensor data. The relevance of using the mathematical apparatus of fuzzy neural networks in the study and calculation of reliability indicators, as well as the development of the NEFClass M fuzzy neural network for calculating the reliability indicators of IoT image sensors is shown. A comparative analysis of Kohonen networks and fuzzy neural networks, as well as the operating principles of these networks, is presented. Kohonen's self-organizing mathematical maps are used for simulation and experimental modeling, traffic forecasting, finding patterns in large data sets and statistical data, identifying patterns of independent and dependent features, compressing and encoding information. The results of using the fuzzy neural subnetwork method in the intelligent analysis of IoT image sensor data are shown in the analysis, experiments and resulting tables. The main advantage of using neural subnetworks is the ability to solve various unstructured tasks and problems.