×

You are using an outdated browser Internet Explorer. It does not support some functions of the site.

Recommend that you install one of the following browsers: Firefox, Opera or Chrome.

Contacts:

+7 961 270-60-01
ivdon3@bk.ru

  • Using Clustering Methods to Automate the Formation of User Roles

    The article solves the problem of automated generation of user roles using machine learning methods. To solve the problem, cluster data analysis methods implemented in Python in the Google Colab development environment are used. Based on the results obtained, a method for generating user roles was developed and tested, which allows reducing the time for generating a role-based access control model.

    Keywords: machine learning, role-based access control model, clustering, k-means method, hierarchical clustering, DBSCAN method

  • Intelligent detection of steganography transform based on containers classification

    The possibility of detection of steganography in digital images based on the classification of stegocontainers is investigated. The obtained results demonstrate the effectiveness of using deep neural networks for solving this problem. The LSB method can be detected using EfficientNet b3 architecture. The achieved classification accuracy is above 97%. Using of steganography methods in frequency domain can be effectively detected by classifying their representation in the form of a digital YCrBr model, with augmentation (vertical and horizontal rotations). The classification accuracy is above 77%.

    Keywords: Steganography, stegocontainer, machine learning, classification, digital image, deep learning, CNN, EfficientNet b3, confidentiality, information protection

  • Extracting key concepts and relationships between them from topical texts in Russian

    The article discusses approaches to solving natural language processing problems such as extracting key concepts or terms, as well as semantic relationships between them to build data-driven IT solutions. The subject of the work is relevant due to the constant growth of volumes of low-structured and unstructured digital text. The extracted information can be used to improve numerous processes: automatic tagging, optimization of content search, construction of word clouds and navigation sections; furthermore, to create draft versions of dictionaries, thesauri, and even bases for expert systems.

    Keywords: natural language processing, term, lemma, semantical relationship, statistical processing, machine learning, word2vec

  • Application of machine learning and natural language processing to develop a recommendation system for the selection of perfumery products

    The paper discusses the use of machine learning in relation to natural language processing (sentiment analysis, semantic proximity analysis) to build a recommendation system for the choice of perfumery products. The topic of the work is relevant in view of the growth of the range of manufactured perfumery products and the complexity of its choice by consumers and promotion by manufacturers. The proposed approaches are relevant for solving this problem due to the accumulated textual reviews and reviews of perfumery products on various websites, including online stores.

    Keywords: machine learning, natural language, sentiment analysis, distributive semantics, word2vec, recommender systems

  • Automated mapping of ORM2-diagram basic elements into OWL2-ontology

    An ontology is a formal, explicit specification of a shared conceptualization of some fragment of domain, which is understandable for both people and intelligent systems. However, it is rather complicated for a domain expert with limited ontology expertise to create complete and consistent ontology, which could answer required competency questions. In this paper we consider the approach with the use of ORM2-diagram as an intermediate model for creating an OWL2-ontology. This approach requires specialized conversion rules allowing to map ORM2-diagram into OWL2-ontology. In this study we revealed that existing conversion rules do not comply with ORM2-semantics. We improved existing conversion rules for Entity Type and Subtyping elements of ORM2 notation. Also we automated the process of mapping ORM2-diagram (consisting of base elements of ORM2 notation - Entity Type, Value Type, Subtyping, Unary Role, Binary Role) into OWL2-ontology as well. As result of this study, we developed a software component, which is part of plugin for ontology editor Protégé. This paper also describes an experiment that confirms the effectiveness of the developed module. It is proven that module allows to exclude mistakes encountered in conversion ORM2-diagram to OWL2-ontology and to reduce conversion time as well.

    Keywords: explicit knowledge representation, visual model, intermediate model, ontology modeling, ontology, OWL2, ontological pattern, ORM (Object-Role Modeling), ORM diagram

  • Ontological Modeling Approach for WHAT-Knowledge by means of ORM2-diagrams and the developed Protege-plugin

    To share and transfer knowledge, they must be presented in an explicit form that is understandable to both humans and computers. The paper proposes an approach to ontological modeling of WHAT-knowledge, which allows representing knowledge simultaneously in two forms: a) in the form of a visual model (ORM2-diagram), understandable to humans, and b) OWL2-ontology, computer understandable. To convert the knowledge representation from one form to another one, it is proposed to use ontological patterns (mapping rules). Currently, there is no software toolkit that allows a) to build an ORM2-diagram and mapping it in the OWL2-ontology, and b) based on OWL2-ontology, build a visual model in the form of an ORM2-diagram. Therefore, we are developing a Protege-plugin, which should provide a) the creation and editing of WHAT-knowledge by building an ORM2-diagram and mapping it in the OWL2-ontology, and b) visualization of WHAT knowledge in the form of an ORM2-diagram, extracting instances of ontological patterns from the OWL2-ontology. The paper provides a functional and structural description of the plugin; examples of its use are given.

    Keywords: WHAT-knowledge, explicit knowledge representation, ontological modeling, ontology, visual model, intermediate model, ontological pattern, ORM2-diagram