(сетевой с №3-2015)
(сетевой с №2-2015)
|Articles and journals | Tariffs | Payments | Your profile|
The use of multilayer neural networks in problems of updating the user interface
Abstract.The subject of the study is the process of forming the parameters of the user interface that is being changed. The author considers the possibility of using neural network methods to process characteristics and classify the categories of user interfaces. To solve the problem studied, the structure of an artificial neural network based on a multilayer perceptron is modeled, the training set of data is prepared and the kit is used to train a simulated neural network. Based on the obtained parameters of the neural network, a mathematical model was constructed and analyzed to solve the problem in question, a study was made of the resulting data for various sets of input data that were not used in training the network. To form the training data set, simulate and train the neural network, the MathLab MathLab system and the Neural Network Toolbox expansion package are used, which allows building a neural network suitable for solving the tasks posed. The conducted research showed that the problem of classifying the parameters of the user interface by a set of input characteristics can be adequately solved with the help of a neural network of direct propagation based on a multilayer perceptron. The most important moments when using this solution are the choice of the architecture of the neural network being created and the preparation of a training data set for building a neural network
Keywords: testing, training, updating, classification, neural network, modeling, user interface, perceptron, Matlab, model
Article was received:19-05-2017
This article written in Russian. You can find full text of article in Russian here .