ABSTRACTS OF ARTICLES OF THE JOURNAL "INFORMATION TECHNOLOGIES".
No. 6. Vol. 23. 2017

To the contents

A. E. Ermakov, Leading Specialist on Computational Linguistics, RCO LLC, Moscow

Semantic Transformations Language for Computer Interpretation of Text

The article is devoted to the development of approach to semantic interpretation of natural language text. The approach was initially described by the author in 2009. The approach is implemented in two stages.
In the first stage, the syntactic parsing is performed for each phrase of the text, producing the network of connected frames. Each frame corresponds to some entity which was mentioned in the phrase of text. The entity could be a single word, a phrase or some name. The frame has attributes whose values correspond to the parts of speech and semantic attributes of the corresponding entity. The frame has a set of slots that contain links to other frames corresponding syntactically subordinated entities in the phrase. Slots are characterized by roles that correspond to the types of syntactic and semantic relationships between entities in the phrase. At the second stage, the lookup of the frames subnetworks is performed under given restrictions. The restrictions on the frames
subnetwork configuration describes the desired situation of the subject area — facts and events. Formally, these configurations correspond to the target rules described by a specially developed formal language — semantic transformations language. These rules allow to add new frames, new frame attributes, as well as new connection between the new and old frames. The semantic transformations language which allows us to describe and interpret arbitrarily complex configurations between frames is described in details. The ability by means of the approach to semantically interpret arbitrarily complex phrase by applying simple rules is illustrated in the article with a lot of examples.
The algorithmically efficient implementation of the semantic interpretator which provides a frames network transformation is described. The algorithm applies the rules by iterations until network is stabilized. The essential feature of the algorithm is in-variance to the order of application of the rules.
The event and fact extraction and sentiment analysis tasks are considered as illustrating examples of the practical applications of this approach.
Keywords: computational natural language processing, computational semantic, semantic representation, formal semantic languages, syntactical analysis of text, semantic interpretation of text, frame network, transformation of semantic structure of phrase, events and facts extraction, sentiment analysis

P. 403–412

V. I. Levin, D. Sc., Professor, e-mail: vilevin@mail.ru, Penza State Technological University, Penza, 440039, Russian Federation
Maximal and Minimal Poly intervals and Optimization in Condition of Uncertainty

The article introduced the concept of comparison of polyintervals, in order to separate larger and smaller polyinterval. The new concept is based on the adopted in the interval mathematics methodology to expand numeric operation to intervals. The main theorem which determines the necessary and sufficient conditions for comparability of polyintervals is proved. An example of using the results of the article to make better decisions in the economy is given.
Keywords: polyintervals, comparison of polyintervals, maximal (minimal) polyinterval

P. 413–417

A. Yu. Romanov, PhD, Senior Lecturer, e-mail: a.romanov@hse.ru, National Research University Higher School of Economics, K. E. Lomotin, Student, e-mail: ke.lomotin@gmail.com, E. S. Kozlova, Student, e-mail: hse.kozlovaes@gmail.com

Machine Learning Methods Application to Automatic Scientific Articles Classification According to UDC
The paper deals with the applicability of modern machine learning methods to the problem of automatic generation of UDC for scientific articles. As the classifiers, such models as artificial neural networks, logistic regression and boosting are considered. Graph algorithms and a prototype software module to generate UDC are designed.
Keywords: text classification, machine learning, artificial neural network, logistic regression, boosting, UDC

P. 418–423

Å. Å. Akimkina, Graduate Student, e-mail: elakimkina@gmail.com, State Educational Institution of Higher Education of the Moscow region University of Technology, Korolev, Moscow region, 141070, Russian Federation

Instrument Approach to the Organization of Data Collection Warehouse Decision Support Systems

Analyzed the problems of modern means of integration in the storage decision support systems A six-level architecture is developed that differs from the widely used three-level architecture in that the levels introduced take into account the problems of extracting data using the latest information technologies from various information resource providers and enable detailed data processing operations within each level. Analytical dependencies for various methods and scenarios of extraction, collection and loading of data are determined; Requirements for the software functionality implementing the six-level service-oriented data storage architecture are formulated.
The approach to organizing data collection in the DDS repository is to select methods and scenarios for collecting, retrieving and loading data based on the unification of the rules of interaction with information resource providers.
This approach allows to solve the problem of data integration of various information resource providers to support decision-making in the Decision Support System (DSS). The use of the developed methods and tools provides an increase in the efficiency of extracting data and loading it into the repository by unifying the processing of electronic sources of various formats regardless of their structure and size.
The developed software functionality allows implementing an instrumental approach to organizing data collection in the DSS repository based on the latest information technologies.
The proposed analytical relationships for various methods of data collection describe the process of integrating the data of the electronic source into the target storage structures.
Keywords: data integration, data collection methods, levels of detail, service-oriented solutions

P. 424–430

A. S. Akopov, Professor, e-mail: aakopov@hse.ru, A. L. Beklaryan, Lecturer, e-mail: abeklaryan@hse.ru, N. K. Khachatryan, Associate Professor, e-mail: nkhachatryan@hse.ru, A. V. Fomin, Associate Professor, e-mail: afomin@hse.ru National Research University Higher School of Economics, Department of Business Analytics, Moscow

The Forecasting System Dynamics of Oil Production using Simulation Modeling

This article presents an original information-analytical system developed using parametric approximation and simulation techniques and designed for scenario forecasting the dynamics of oil production in wells. This system is implemented as a software, the core of which is a simulation model developed in an IDE AnyLogic, integrated with a database and describes the dynamics of production at each well, taking into account implemented and planned geological and technical measures (GTM). A prototype system has been successfully implemented in Russia's largest oil company and used to predict the incremental oil production around the pool of all existing wells (more than 100,000 wells in the ten-year time frame), as well as assessing the potential effect of GTM for subsequent reallocation of resources between wells.
Keywords: simulation modeling, parametric approximation, oil production forecast, service-oriented solution, AnyLogic

P. 431–436

G. G. Bulychev, DSc., Professor, e-mail: geo-bulyehev@mail.ru Moscow Technological University, Moscow

Simulation of Dynamics of the Hollow Cylinder by the Method of Spatial Characteristics

The method of spatial characteristics is a numerical and analytical technique in which the finite difference relations built after bringing the original model of the dynamics of deformed solid body to the characteristic form. From a programmer's point of view, this method can be classified as direct explicit grid-characteristic method of the first or second order approximation. The first order approximation of the remainder term has a parabolic shape, so spurious vibrations caused by the hyperbolicity of the original equations, damped. When considering the advantages and disadvantages of this method mentioned earlier that one of the main obstacles to its wider dissemination, is the difficulty of constructing algorithms for objects with curved surfaces. For such bodies spatial grid on which modeling check ditsya is curved and contains a different size of the cell that is not valid for the finite approximation templates. In this article, on the example of the dynamics of a hollow cylinder, it is shown that this difficulty can be overcome, but with some limitations and not for any classes of problems using divergence of the differential equations characteristic parts. To a first order approximation it was constructed numerically, then was realized and tested the corresponding algorithms. Algorithms showed good accuracy and stability in a large time interval. Method applied here can be used to study the dynamics and destruction of various dynamic bodies with curved surfaces.
Keywords: a method of spatial characteristics, numerical simulation, dynamics of constructions, algorithms

P. 437–442

I. V. Mashkina, Professor, e-mail: mashkina.vtzi@gmail.com, A. O. Kupritanov, Undergraduate, e-mail: anton_mail@ro.ru Ufa State Aviation Technical University (UGATU), Ufa, Russia

The Classification Scheme and Modern Viruses Models

The classification scheme of the modern-day viruses software is proposed in this paper. The proposed scheme is multi-layer, the following bases of classification are used in the first layer: static, dynamic and combined. Static virus has a permanent body and can't modify it, such code has a signature and is easily detected by antivirus programs. Dynamic virus don't have a permanent body it can modify its code itself, using several techniques (encryption, obfuscation) to make more difficult automatic analysis by antivirus programs. Event-driven process chains (EPC) are used when developing the virus models based on the suggest classification. These models appropriate to describe the virus architecture in detail and show action logic, functions, events, and resources of virus. The proposed classification scheme and models of modern viruses are required to design an adequate mode of antivirus software. Appling of the models of malware samples is appropriate when designing antiviruses to prevent infection of the computer systems.
Keywords: malware, virus, virus classification scheme, viruses model, static virus, dynamic virus, combined virus

P. 443–448

R. S. Zagidullin, Dr. Associate Professor, e-mail: zag@bmstu.ru, A. S. Chernikov, Head of Educational & Research Laboratory, e-mail: chernik@bmstu.ru, A. A. Chibisov, Programmer of the Research & Educational Centre, e-mail: chibisov.alexandr.25@gmail.com,
Moscow, 105005, BMSTU

Integration of Moodle and Electronic University Systems at BMSTU

The article considers a method of constructing integrated solutions for remote laboratory practicum. Described the remote network connections technologies; management and monitoring techniques of laboratory equipment; increase computers usage effi­ciency. Reviewed a sample implementation of a remote laboratory practicum on electronics for the BMSTU students.
Free platform MOODLE was integrated with closed information control system Electronic University (EU) of Bauman Moscow State Technical University which serves for support of administrative work on control of educational process. The following main problems were solved:
• creation of unified data representation in the two systems;
• creation of students' and training courses' databases in Moodle based on data from EU.
The Moodle data storage format was brought in line with the requirements of the EU system. Unique software was developed.
Combination of Moodle with EU gives new quality — all stages of University activity join together: teaching, learning and administrative control of educational process. This concept can be used by any University to reduce "human factor" errors and to increase performance.
Keywords: Information control system, MOODLE, database, Electronic University, student, course structure, grade, human factor

P. 449–456

N. I. Yusupova, DSc., Professor, e-mail: yussupova@ugatu.ac.ru, O. N. Smetanina, DSc., Professor, e-mail: smoljushka@mail.ru, M. M. Gayanova, PhD, Associate Professor, e-mail: maya.gayanova@gmail.com, Ufa State Aviation Technical University, Ufa

Models and Methods of Information Processing in the Analysis of Educational Programs (on the Example of Training Computer Scientists)

The results of many years researching by the authors on the analysis of educational programs are presented to identify the positive experience of training IT professionals in leading foreign universities in this field (Lausanne Higher Polytechnic School, Osaka University, Stanford University, Karlsruhe Technical University, etc.). Another goal of the comparative analysis of educational programs is the possibility of forming educational routes within the framework of academic mobility.
The general features of educational programs for the training of computer scientists are reflected, both in the content aspect and in the form ofpresentation. It is noted that the differences in the form (structure, format (Excel, PDF, HTML), presentation language, etc.) and significant volumes of the collected documentary base make the procedure for analyzing educational programs a difficult and complex task (evaluating the semantic proximity of keywords; finding a substring that satisfies the pattern of regular expression; map­ping the results of comparative analysis of educational programs in the form of knowledge representation models to provide solutions; identifying similar structures of educational programs; adaptation of ready solutions to a similar situation on the basis of reasoning based on precedents, etc.). It is revealed that some of the problems are related to the processing of weakly structured data.
The results of the educational programs analysis by other researchers are given, the models and methods used for analysis are noted. Among the analysis the following tools are revealed: methods of expert assessments, qualitative content analysis, clustering methods, methods of searching for associative rules, etc.
The results of processing educational programs information for the training of computer scientists are presented to identify their similar structures using clustering methods. The results were obtained using the analytic platform Deductor. For students in higher education institutions, whose educational programs have similar structures (they are in the same cluster), an educational route with participation in academic mobility programs can be more comfortable.
The possibility of using the results of comparative educational programs of Russian universities to support networking of universities is noted.
Thus, the issues of analytical processing of information about educational programs are related to the problems of processing poorly structured data. To solve this problem, special models and methods are used, usually based on ideas of artificial intelligence and knowledge engineering.
The possibility of using the results of the comparative analysis of educational programs of Russian universities is noted to support network interaction of universities.
Keywords: data processing, structure of the educational program, cluster analysis, neural networks, weakly structured data

I. L. Gavrilova, Senior Lecturer, e-mail: igavr@grsu.by,
Yanka Kupala State University of Grodno, Belarus, Grodno

Experience in Development and Implementation of the Learning Process of the University Electronic Educational Complex

Administering to the subject of research — process of training of students in conditions of informatization of educational process. General provisions are considered.
The purpose of the study is to show experience in the development and implementation into study electronic educational complex on discipline "Antenna-feeder devices". Relevance of the research is concluded in using common technical tools in the educational process.
Experience in the development and implementation into study electronic educational complex on discipline "Antenna-feeder devices" is shown in the main part of the work. Main elements of the complex are described in the beginning of the work, structure is also shown. Content of the main parts are disclosed. Modular approach in complex organization is viewed. Features of the technical training are marked.
Special attention is paid for creature specialized sector for working out virtual laboratory which applies to practical part of the complex. Components of the virtual laboratory are described in the work. Universality of the designed industry is considered for conduction practical classes in any discipline and it application.
One of the main element of the complex is designing computer system of testing knowledge. One of the advantage of this program is diversity of the tasks and also using in testing tasks graphic, sound, videomaterials.
In section of the control of knowledge in the complex main role drained to conduction controlled work of students. Experience of the desiglned work is shown. Different shapes of conduction classes are shown.
Advantages of the designed complex are described in the conclusion.
Received results of the study afford formulate requirements to the content of the electronic complex and can be used in elaboration new methodical complexes in university educational process.
Keywords: methodical complex, e-learning, development environment, independent work, educational process, testing, remote education

P. 464–468

A. E. Saak, D.Sc., Associate Professor, Department of State and Municipal Administration, Head of Department, e-mail: saak@tgn.sfedu.ru; Southern Federal University, Rostov-on-Don — Taganrog, Russian Federation

Sheduling of Sets of Parabolic Tasks in Grid Systems

The effectiveness of Grid systems functioning is mainly based upon the scheduling policy realizing by the resource management and job scheduling system. Grid systems with centralized architecture, which consist of number of sites containing parallel systems with multisite scheduling that is characterized by the possibility of multiprocessor task performance on several sites simultaneously, are basically modeled by resource quadrant. Incoming user's task is characterized by two parameters: by number of the required processors and the action's execution time. Incoming user's task, being actually monitoring by the dispatcher, is modeling by the resource rectangle with the horizontal and vertical dimension, which are correspondingly equaled to the number of resource units of time and quantity of processors required for task performance.
Optimal redistribution of the calculation and timing recourses basically has an exponential labor intensity, as well as an heuristics algorithms of polynomial labor intensity are the things of practical importance. The author suggests resource rectangles environment as methodology of polynomial completeness in scheduling theory. The heuristic polynomial algorithms for processor and time resource distribution, based upon the introduced operations on resource rectangles, are suggested in resource rectangles environment.
Scheduling quality is estimated by the author as heuristic measure which reaches its minimum at in-square packing without emptiness. The problem posed in the paper concerns polynomial algorithms adaptation for precise form sets of parabolic type tasks.
Leveled polynomial algorithms of scheduling are recommended to be put in use in dispatching of the parabolic type access in Grid systems of the precise form and structure.
Keywords: Grid systems with centralized architecture, multisite mode of task processing, recourse rectangle, an area of recourse rectangles, quadrant type task, parabolic type task, non-Euclidean heuristic measure, polynomial completeness of an algorithm, polynomial algorithms for processor and time resource distribution, leveled algorithms in height, leveled algorithms in length

P. 469–475

V. I. Sumin, Professor, e-mail: viktorsumin51@yandex.ru, Voronezh institute of the Russian Federal Penitentionary Service, T. E. Smolentseva, Associate Professor, e-mail: smoltan@bk.ru, Lipetsk state technical University

The Formation of the Structure of Hierarchical Multi-Level Organizational Systems

The article describes the process of building the structure of organizations and their management systems, which are multilevel in nature. The control system of such organizations are seen as complex hierarchical multi-level organizational system. In the article the approach to formation of the newly created organization structure in a hierarchical multi-level organizational system implemented on the basis of questionnaires, expert and cluster analysis. With the help of the cluster analysis are determined by structural hierarchical elements of such an organization. In the future, the structure of the organization is represented as a directed graph without contours. The structure of the graph is constructed in accordance with the hierarchy of objectives of the organization.
Keywords: hierarchical multi-level organizational system, management system, questionnaire survey, expert analysis, cluster analysis, directed graph, an iterative method

P. 476–480

To the contents