UDC 007:681.512.2
NEURAL NETWORKS FOR USER IDENTIFICATION BASED ON NEWS SITE VISITS ANALYSIS
Yu. Kashirin, Dr. Sc. (Tech.), full professor, RSREU, Ryazan, Russia;
orcid.org/0000-0003-1694-7410, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
The article deals with the design of neural networks with knowledge. The designed networks are intended to solve the problem of classifying users according to their psychosomatic type. An original method of building networks using ICF-ontologies, structured in accordance with the causal taxonomy of the information needs of the Web site user, is presented. To design an ontological causal taxonomy, elements of the theory of hierarchical numbers are used, which allow calculating the semantic similarity of ontology concepts for its optimization. The features used in the taxonomy are applied later in the formation of a set of basic features (S-elements) of neural network. Software implementation of the approach proposed uses the subject area of current news site with many multidirectional information headings. The purpose of using a neural network is to classify users according to their psychosomatic type based on the analysis of visits to site materials by various users. As a practical part, experiments planned and delivered within the framework of Python v.3 software toolkit (Anaconda 3) are considered. The results obtained allow us to positively assess the effectiveness of the technology proposed. The aim of the work is to create an original method for constructing neural networks for classification problems using causal taxonomies in ICF ontology format using hierarchical numbers.
Key words: neural networks, causal taxonomy, psychosomatic user identification, ICF ontology, hierarchical numbers, semantic proximity.