Development of Architecture and Software Implementation of Deep Neural Network Models Repository for Spatial Data Analysis

Abstract

The article presents the key aspects of designing and developing a repository of deep neural network models for analyzing and predicting the development of spatial processes based on spatial data. The framework of the system operates on the basis of the MVC pattern, in which the framework is decomposed into modules for working with the system's business logic, its data and graphical interfaces. The characteristics of the developed web interfaces, a module for visual programming for editing of models, an application programming interface for unified interaction with the repository are given. The stated aim of the study determined the structure of the scientific article and the results obtained. The paper describes the functional requirements for the repository as a signific part of software design, presents the developed formalized storage scheme for neural network models, describes the aspects of development of a repository of neural network models and an API of the repository. The developed system allows us to approach the solution of the scientific problem of integrating neural networks with the possibility of their subsequent use to solve design problems of the digital economy.