The scientific problem of improving the efficiency of the process of developing adaptive information systems (AIS) is considered: decrease in economic costs and complexity of software implementation, increasing adaptability, quality and productivity of work. The existing methodologies for the development of information systems are not focused on solving the problems of automating information processing processes, adapting program modules, which does not allow solving the problem with their use. Analysis of approaches to solving these subtasks showed the possibility of effective application of neural network technologies. Therefore, this paper considers the development and development of theoretical foundations and software tools based on machine learning technologies and neural networks, combined within the framework of the methodology of structural-parametric synthesis of AIS. The scientific novelty of the methodology lies in the use of a neural network architecture and the addition of a stage for the implementation of neural network components based on new neural network methods. The principles of the methodology are formulated, its general structure is formalized and the decomposition of the main stages is carried out. Approbation of the methodology in two subject areas made it possible to reduce economic costs and complexity of software implementation, increase the adaptability and quality of AIS. The results obtained can be used to organize the development process of various information systems. Further research consists in approbation of the methodology and neural network methods in new subject areas.

Authors: A. D. Obukhov

Direction: Informatics, Computer Technologies And Control

Keywords: Adaptive information systems, neural network technologies, machine learning, methodology of structural-parametric synthesis, methods of analysis and information processing

View full article