Using the multilayer neural network for tasks of updating the user interface

Abstract

This article presents issues related to model building for tasks of forming and modifying a custom user interface. We describe the possibility of using neural network methods to process the input data and classify the categories of user interfaces. The investigated task is solved by modeling the structure of an artificial neural network based on a multilayer perceptron. The preparation of a training data set and the use of this set for the training of neural network are provided. Input layer is considered a distributor of the signals from the external actions of network users with their interfaces. The output layer is considered a generator of some structural parameters of the user interface. Using the obtained parameters of the neural network the mathematical model constructed and analyzed for solving the above problem. The resulting data tested for various input data sets that were not used for training the network. In this paper, we investigate the possibility of applying the designed model for processing information needs of users, analysis and accumulation of the various parameters of user interface, and therefore applicability of this model for forming the most suitable solutions for users. © SGEM2017. All Rights Reserved.