Conference PaperPDF Available

Epidemic effects in network industries

Authors:

Abstract and Figures

Modern research in the field of global cellular telephony says that technology reveals the effects of an established base in acceptance (of a new technology) and that cellular diffusion varies between countries and groups of countries due to technological, socioeconomic and regulatory factors that influence the diffusion process. The effects of an established base may come from the extra effects of social dissemination, such as social learning through uncertainty and regulation pressures. With a severe decrease in the intensity of use, the epidemic effect is largely absent or at least overshadowed by other factors, notably consumer heterogeneity. Hence it follows that the driving forces behind diffusion are constantly falling prices and / or increasing quality of the network services.
Content may be subject to copyright.
Ministry of Education and Science of Ukraine
National Aviation University
Software Engineering Department
International Conference on
Software Engineering
April 12 14, 2021
Proceedings
Kyiv 2021
International Conference on Software Engineering
2
Conference organizers and partners
National Aviation University (Kyiv, Ukraine)
www.nau.edu.ua
Software engineering department
www.sed.nau.edu.ua
Artificial Intelligence and Knowledge
Engineering Research Lab (Cairo, Egypt):
http://aiasulab.000webhostapp.com/Ambassadors.html
The Association of Developers
and Users of Intelligent Systems
http://www.aduis.com.ua/
Faculty of Computer and Information Sciences,
Ain Shams University (Cairo, Egypt):
http://cis.asu.edu.eg/
The Institute for Information Theories
and Applications FOI ITHEA® (Bulgariya)
http://www.ithea.org/index.html
ITHEA®
International
Scientific
Society
International Scientific Society
ITHEA® ISS
http://www.ithea.org/iss/iss.html
V. M. Glushkov Institute of Cybernetics of the NAS
of Ukraine (Kyiv, Ukraine)
http://incyb.kiev.ua/?lang=en
International Conference on Software Engineering
3
NATIONAL AVIATION UNIVERSITY
(www.nau.edu.ua)
For almost 80 years of its history more than 200,000 highly skilled professionals have
been trained in this higher aviation educational institution. Among them there are well-
known scientists, heads of aviation companies, enterprises, organizations and institutions
that provide aircraft flights, their maintenance and repair, transportation of passengers
and cargo.
Strong scientific and pedagogical schools enable to train not only engineers but also
economists, lawyers, environmentalists, translators, psychologists, sociologists and
others.
SOFTWARE ENGINEERING DEPARTMENT
www.sed.nau.edu.ua
Software engineering department studies more than 900 students. Students obtain
fundamental knowledge and practical skills in software development area as well as
theoretical basic in software development life-cycle management.
AIN SHAMS UNIVERSITY AI AND KNOWLEDGE ENGINEERING LAB
http://aiasulab.000webhostapp.com/
The Ain Shams University Artificial Intelligence and Knowledge Engineering Lab was
founded in 2005 and belongs to the Computer Science department, Faculty of Computer
and Information Sciences, Ain Shams University, Cairo, Egypt.
International Conference on Software Engineering
4
The Lab is currently composed of 3 full Profs., 3 Assistant Prof., 15 Drs (PhD holders) 5
of them from Arabic countries, 10 PhD students and 5 Msc students.
Laboratory is working in the following disciplines:
Intelligent Information Processing, Data Mining.
Knowledge Engineering and Expert Systems.
Image Processing and Pattern Recognition.
Data Science and Big Data Analytics
Intelligent Education and Smart Learning.
Medical Informatics, Bio-informatics and e-Health.
Computing Intelligence and Machine learning
Internet of Things (IoT),
Technologies for HealthCare;
Space Science and Cosmic Rays Big Data Analytics.
In addition, the lab includes 17 scientific international ambassadors from Germany,
Canada, Italy, Georgy, Ghana, Cyprus, Ucraine, Romania, Greece, Japan, Mexico, and
Belgium
THE ASSOCIATION OF DEVELOPERS AND USERS
OF INTELLIGENT SYSTEMS
http://www.aduis.com.ua/
Association of Developers and Users of Intelligent Systems (ADUIS) consists of about
one hundred members including ten collective members. The Association was founded in
Ukraine in 1992. The main aim of ADUIS is to contribute to the development and
application of the artificial intelligence methods and techniques.
Association has long-term experience in collaboration with teams, working in different
fields of research and development. Methods and programs created in Association were
used for revealing regularities, which characterize chemical compounds and materials
with desired properties.
International Conference on Software Engineering
5
FACULTY OF COMPUTER AND INFORMATION SCIENCES, AIN SHAMS
UNIVERSITY (Cairo, Egypt):
http://www.shams.edu.eg/598/news#
Faculty of computer and information sciences was founded in 1996 and consists
of 5 departments, namely; Computer Science, Information Systems, Computers
Systems, Scientific Computing, and Basic Sciences.
Faculty of computer and information sciences announces it's distinct programs by credit
hours system:
Software Engineering;
Artificial Intelligence:
Cyber Security;
Digital multimedia.
Faculty of computer and information sciences also provides Bioinformatics program
which accepts Thanawya Amma science section
INSTITUTE FOR INFORMATION THEORIES AND APPLICATIONS FOI ITHEA®
(BULGARIA)
http://www.ithea.org/
The Institute for Information Theories and Applications FOI ITHEA® is an international
nongovernmental organization functioning since 2002 year. ITHEA® is aimed to support
international scientific research through international scientific projects, workshops,
conferences, journals, book series, etc.
International Conference on Software Engineering
6
INTERNATIONAL SCIENTIFIC SOCIETY
ITHEA® ISS (INTERNATIONAL SCIENTIFIC SOCIETY)
http://www.ithea.org/iss/iss.html
ITHEA® International Scientific Society (ITHEA® ISS) is aimed to support growing
collaboration between scientists from all over the world.
Till now, the ITHEA® International Scientific Society has been joined by more
than 4000 members from 53 countries from all over the world:
ITHEA® Publishing House (ITHEA® PH) is the official publisher of the works of the
ITHEA®ISS. The scope of the issues of the ITHEA® ISS covers the area of Informatics
and Computer Science. ITHEA® PH welcomes scientific papers and books connected
with any information theory or its application.
ITHEA® ISS has four International Journals, established as independent scientific printed
and electronic media, published by ITHEA® Publishing House:
International Journal “Information Theories and Applications” (IJ ITA), since 1993;
International Journal “Information Technologies and Knowledge” (IJ ITK), since
2007;
International Journal “Information Models and Analyses” (IJ IMA), since 2012;
International Journal “Information Content and Processing” (IJ ICP), since 2014.
Every year, ITHEA® ISS organizes many scientific events called “ITHEA® ISS Joint
International Events of Informatics" (ITA). ITHEA® ISS has
organized 135 International Conferences and workshops and has published 45 scientific
books: monographs, thematic collections and proceedings.
All publications of ITHEA® ISS are accessible freely via ITHEA® main site www.ithea.org
as well as via ITHEA® Digital Library (IDR) http://idr.ithea.org.
International Conference on Software Engineering
7
The great success of ITHEA® International Journals, Books, and Conferences belongs to
the whole of the ITHEA® International Scientific Society. Membership in ITHEA® ISS is
free.
V.M.Glushkov Institute of Cybernetics of the NAS of Ukraine
http://incyb.kiev.ua/?lang=en
M.Glushkov Institute of Cybernetics of the NAS of Ukraine is a Scientific Centre widely
known in Ukraine and beyond it engaged in the solution of fundamental and applied
issues of informatics and computer science as well as introduction of their methods and
means to various spheres of human activities.
Today the main areas of the research of the Institute are:
Development of the general theory and methods of system analysis, mathematical
modelling, optimization and artificial intelligence;
Development of the general theory of management, methods and means of arrangement
of intelligent control systems of different levels and purposes
Creation of the general theory of computing machines and development of perspective
computing equipment, artificial intelligence and informatics;
Creation of perspective systems of mathematical support of general and applied
purposes;
Development of new information technologies and intelligent systems; and
Solving of fundamental and applied problems of informatization of the society.
Scientific schools of computer mathematics and discrete optimization, mathematical
theory of computing systems and artificial intelligence, system analysis and stochastic
programming, mathematical theory of reliability and programming theory were forming for
fifty years at the V.M.Glushkov Institute of Cybernetics of the National Academy of
Sciences of Ukraine and they gained world’s recognition.
International Conference on Software Engineering
8
Conference topics
Theoretical Basics of Software Engineering
AGILE and Model-Driven Approaches to Software Development.
Approaches to Software Development Life Cycle Processes Improvement;
Business Process Management and Engineering
Empirical Software Engineering;
Formal Foundations of Software Engineering;
Frameworks and Middleware;
Model-Driven Engineering;
Software Development Technologies;
Software Engineering Standards;
Software Patterns and Refactoring;
Applied Aspects of Software Engineering
Approaches and Models to Estimate Expenses of Future Software Project;
Component-Based Software Engineering;
Requirements Engineering;
Service-Oriented Software Engineering;
Software Designing;
Software Quality;
Software Maintains and Evolution;
Software Testing.
Software Engineering Application Domains
Agent and Multi-Agent Systems;
Biotechnologies and Smart Health Technologies;
Cloud Computing;
Data Bases and Knowledge Bases;
Image Processing and Computer Vision;
International Conference on Software Engineering
9
Internet;
Security;
Software Development for Mobile Operation Systems.
Real-Time Systems;
Network and Data Communications;
User-Centered Software Engineering.
International Conference on Software Engineering
10
Steering committee
Abdel-Badeeh M. Salem (Ain Shams University, Egypt);
Albert Voronin (National Aviation University, Ukraine);
Krassimir Markov (Institute of Information Theories and Applications, Bulgaria);
Olena Chebanyuk (National Aviation University, Ukraine);
Serhii Zybin (National Aviation University, Ukraine);
Vitalii Velychko (V.M.Glushkov Institute of Cybernetics of National Academy of
Sciences of Ukraine).
Program committee
Antinisca Di Marco (University of L'Aquila, Italy);
Besik Dundua (Tbilisi State University, Georgia);
Eugene Machusky (National Technical University of Ukraine Igor Sikorsky Kyiv
Polytechnic Institute, Ukraine);
George Totkov (University of Plovdiv, Bulgaria);
Krassimira Ivanova (University of Telecommunications and Post, Bulgaria);
Michael Gr. Voskoglou (Graduate Technological Educational Institute of Western
Greece School of Technological Applications, Greece);
Mohamed Ismail Roushdy (Future University, Egypt);
Mykola Guzii (National Aviation University, Ukraine);
Nagwa Badr (Ain Shams University, Egypt);
Sedat Akleylek (Ondokuz Mayıs University, Turkey);
Vladimir Talalaev (National Aviation University, Ukraine).
Organization committee
Mykyta Krainii (National Aviation University, Ukraine);
Tetiana Konrad (National Aviation University, Ukraine);
Yana Bielozorova (National Aviation University, Ukraine);
Yuliia Bezkorovaina (National Aviation University, Ukraine).
International Conference on Software Engineering
Table of contents
SECTION: THEORETICAL BASICS OF SOFTWARE ENGINEERING .......................... 13
CONCEPT OF DATA SETS REUSE IN SOFTWARE PRODUCT LINE APPROACH
Olena Chebanyuk .............................................................................................................. 13
NORMALIZATION OF LIFE CYCLE MODELS IN TASKS PLANNING SOFTWARE
PROJECTS
Vladimir Talalaev, Larysa Postavna .................................................................................. 17
AUTOMATIC ONTOLOGY CREATION FROM NATURAL LANGUAGE TEXT
Anna Litvin, Vitalii Velychko, Vladislav Kaverinsky ........................................................... 21
APPLICATION FEATURES OF AGILE APPROACHES IN SOFTWARE ENGINEERING
Iryna Gruzdo, Ihor Michurin ............................................................................................... 26
AN APPROACH OF DESIGNING RESTFULL API MODEL FOR DISTRIBUTED WEB
APPLICATIONS
Yurii Milovidov .................................................................................................................... 30
SECTION: APPLIED ASPECTS OF SOFTWARE ENGINEERING ................................. 33
THE METHOD MULTICRITERIAL EVALUATION OF VERIFY THE QUALITY
PROGRAM SYSTEMS IN USE
Yuliia Bezkorovaina............................................................................................................ 33
International Conference on Software Engineering
12
SOFTWARE TECHNOLOGY OF FLOOD SIGNALING AND FORECASTING IN THE
TRANSCARPATHIAN REGION
Volodymyr Polishchuk ........................................................................................................ 42
SECTION: SOFTWARE ENGINEERING APPLICATION DOMAINS .............................. 47
AUTOMATIC SEARCH FOR INVOLVANTS OF INFORMATION MESSAGES IN THE
VOICE DATABASES
Serhii Zybin, Yana Bielozorova.......................................................................................... 47
BAYESIAN REASONING AND MACHINE LEARNING
Michael Gr. Voskoglou ....................................................................................................... 51
UKRVECTŌRĒS: AN NLU-POWERED TOOL FOR KNOWLEDGE DISCOVERY,
CLASSIFICATION, DIAGNOSTICS & PREDICTION
Vitalii Velychko, Kyrylo Malakhov, Oleksandr Shchurov ................................................... 57
MULTILINGUAL TEXT CLASSIFICATION USING TRANSFORMER APPROACH
Larysa Chala, Mykola Dudnik, Mariia Sokolovskaya ........................................................ 64
EPIDEMIC EFFECTS IN NETWORK INDUSTRIES
Gorbachuk Vasyl, Dunaievskyi Maksym, Syrku Andrii ..................................................... 68
International Conference on Software Engineering
SECTION: THEORETICAL BASICS OF SOFTWARE ENGINEERING
CONCEPT OF DATA SETS REUSE IN SOFTWARE PRODUCT LINE APPROACH
Olena Chebanyuk
Abstract: Approach of data reuse to estimate reliability of software components before
adding them to repository is proposed in this this paper.
Keywords: Software Product Line, Repository of Software Components, Systematic
Reuse.
Introduction and State of Art in Reuse Problem
Software Product Line is perspective and modern approach focused on systematic
software development artifacts reuse. Modern tendencies and approaches are focused
mostly on reuse of software modules and design solutions.
Following this tendency one of the perspective and practical aspects of Software Product
Line approach is designing of software components repositories. From the theoretical
point of view, such repositories simplify the procedure of software development, namely
such activities as software construction, designing and development (IBM, 2016).
When new software module of project is added to repository of software components
the actual question about preliminary software quality estimation becomes an actual.
Importance of this question is explained by the next factors: IT company, then customer,
then software user must be sure that functionality is realized correctly, and hidden
mistakes of execution of software bugs will not fail the correct execution of software.
Modern approaches are focused on new testing techniques usage (Chebanyuk and
Palahin, 2016), for example Test-Driven Development (TDD). Traditionally TDD is
involved in forward engineering, when data sets (test cases and test suites) are prepared
before realizing project features. Preparing data and expectation of needed results in
International Conference on Software Engineering
14
forward engineering allows increasing the speed and accuracy of the process selecting
from variety of software modules those, which satisfies to the requirements of the new
software project in the best way.
Contribution of this thesis: this thesis proposes general idea of software modules quality
estimation approach that considers data sets reuse. Data sets are test cases and test
suites that are accumulated from previous projects in software product line approach.
Task to propose a concept of data sets reuse in software product line approach for
estimating software components quality.
The essence of the proposed approach is represented in the figure 1.
view
models
1
JSon Data
Schemaes
Problem
Domain
Tree
Data
Sets
2.1
JSon Data
Schemaes 3
5
4
2.2
Figure 1. General scheme of the data sets reuse in product line approach
Reuse concept is based on the following actions:
1. Problem domain tree is designed (Chebanyuk and Palahin, 2019). Data sets that
are represented as test suites are gathered from the previous projects, performed
in software company. They are matched to problem domain tree processes.
International Conference on Software Engineering
2. View models of MVC project are investigated (Chebanyuk and Palahin, 2018).
Then JSON (or XML) schemas, representing data structures, namely project Data
Models are extracted (message 1 in the figure 1). Good example of data
structures processing representing is proposed in book (Seriy at el,, 2016).
3. If project topic matches to problem domain processes (message 2.1 in the figure
1) request to Software Product Line repository of data sets is performed (message
2.2. in figure 1).
4. The closest data schemas to existing View Models are returned (message 3 in
figure 1)
5. Then data for automatic testing are refined (Message 4 in figure 1). Consider
possible example of such a refinements: in View Model five data fields are
represented. The closest data scheme contains four fields that matched to View
Model. Request can fill extra data field by default values (for example “nullor
zero) or perform a search in other data schemes field that is match to View Model.
Then missing field is merged with existing data model.
6. Automatic test are executed (Chebanyuk and Palagin, 2016). Then the request to
extort test results to components repository is performed. Performing similar
procedures (messages 3 and 4 in figure 1) testing results are extracted from the
repository and the further comparison operation of expected and obtained results
is performed.
Bibliography
1. Chebanyuk O. & Palahin O. (2016) Model-Driven Test Approach in Agile Based on
User-Stories Analysis. International journal “Informational Technologies and
Knowledge”, Vol 10, number 4, 2016, 303-316
2. Chebanyuk O. & Palahin O. (2018) A multi-layer approach of view models designing.
International journal “Informational Content and Processing”, 2018, 203-216.
3. Chebanyuk O. & Palahin O. (2019) Domain Analysis Approach. International journal
“Informational Content and Processing”. Volume 6, Number 2, 2019, 3-20.
International Conference on Software Engineering
16
4. Alex Seriy, Bhargav Perepa, Christian E. Loza, Christopher P. Tchoukaleff, Gang
Chen, Ilene Seelemann, Kurtulus Yildirim, Rahul Gupta, Soad Hamdy, Vasfi Gucer
(2016) Getting Started with IBM API Connect: Scenarios Guide, 1-150.
Authors' Information
Olena Chebanyuk D.Sc., Professor of Software Engineering
Department, National Aviation University, Kyiv, Ukraine, e-mail:
chebanyuk.elena@ithea.org
Major Fields of Scientific Research: Model-Driven Architecture, Model-
Driven Development, Software architecture, Mobile development,
Software development
International Conference on Software Engineering
NORMALIZATION OF LIFE CYCLE MODELS IN TASKS PLANNING SOFTWARE
PROJECTS
Vladimir Talalaev, Larysa Postavna
Abstract: An approach to solving the problem of choosing a life cycle model for software
project management and a step-by-step algorithm for its implementation are proposed.
The approach is based on the application of the procedure of normalization of life cycle
models using a structural template, as well as the implementation of the procedure of
expert-analytical comparison of normalized models with the profiles of the context of the
software project.
Keywords: Software Engineering, Life Cycle Models, Software Project.
Introduction
The success of any software development project can be assessed within a set of
metrics agreed with key stakeholders, among which the three-dimensional metric "quality
- cost - time" occupies a dominant position. Also known as the "iron triangle" problem,
this metric allows you to formulate a general criterion for the effectiveness and success
of the project. A brief definition of this criterion can be summarized as follows: “A
software project can be considered successful if such static and dynamic project
parameters are selected and implemented that improvement of any parameter of the
“iron triangle” sides cannot be achieved without losing the balance of interests of any of
them key "stakeholders of the project". The implementation of this approach involves the
presence in the arsenal of project participants at least a complete set of project
characteristics, in which the processes of local optimization of the multicriteria task of
selecting the best by the given criterion give the desired effect in achieving global
optimum. Among this set of project characteristics, an important role is given to the
choice of life cycle model. This is one of the key decisions determines the optimal choice
of structural and functional parameters and characteristics of all processes of the project
International Conference on Software Engineering
18
life cycle, and, consequently, significantly affects the achievement of the optimal global
criterion of
The following solution to the problem of choosing a life cycle model for a software project
is based on the concept of "life cycle model" in the narrow sense, when the main
processes of the model are only the processes of the software product development
stage. This approach is due to the widespread use in software engineering practice of
descriptions of typical life cycle models, which are limited to the stage of development. Of
course, this approach narrows the scope of the results, but usually leads to more
adapted to the practical application of research results.
In the problem under consideration, the life cycle model is defined as a description of a
pattern (template) that summarizes a particular practice of software development and
reproduces the key properties of life cycle processes in structural-functional, temporal,
spatial and behavioral projections. Since the basis of software projects are the processes
of software development (design, construction, implementation of testing, etc.), the
correct choice of life cycle model is crucial in the planning of software projects.
Analysis of information sources on software project planning, in most best practices, the
decision to choose a life cycle model is chosen by project team management based on
their experience and intuition by analytically comparing the context of software
development with information on the advantages and disadvantages of a known life
model cycle. Along with the subjectivity of this choice, there is a significant dependence
on the completeness of the knowledge base for best design practices, as well as a
significant dimension of the problem of analytical comparison of project context profiles
with descriptions of life cycle model options. The synergistic effect of these negative
factors significantly impairs the effectiveness of design decisions and increases their risk.
Reducing the risk of life cycle model selection procedures and reducing the undesirable
consequences of their impact is possible through their formalization and further
automation within the use of software CASE-design tools.
International Conference on Software Engineering
In the proposed approach, a step-by-step selection algorithm is implemented, the main
operating blocks of which are:
1. Formation of a template for normalization of the description of models of a life cycle of
software products.
2. Collection of information on best practices in the use of life cycle models in the
practice of software development and systems.
3. Normalization of life cycle models from the knowledge base in accordance with the
created template and the formation of a database of models.
4. Development of a structural template to describe the context profile of the software
project.
5. Implementation of expert-analytical procedure for comparing the profile of the software
project with normalized life cycle models. Formation of a set of dominant variants of life
cycle models for selection of the best.
6. Application of the expert procedure of the final choice of the life cycle model and its
application for planning the working structure of the project.
Fig. 1 Templates for normalizing life cycle models and building context profiles of
software projects
International Conference on Software Engineering
20
Figure 1 shows the structure of life cycle model normalization templates and the context
profile structure, according to which more than 20 descriptions of typical life cycle models
were processed and a database for software projects was created.
Conclusion
The application of the proposed approach, as well as its implementation in the form of an
appropriate step-by-step selection algorithm can significantly improve the process of
planning software projects, reduce the risks of their implementation and create a strong
knowledge base for effective software project management.
Bibliography
1. Gruber T.R. Toward Principles for the Design of Ontologies Used for Knowledge
Sharing // International Journal Human-Computer Studies. 1995. Vol. 43(56).
P. 907928.
2. Shamova TI, Davydenko TM, Rogacheva NA Adaptive school: problems and
prospects. - Arkhangelsk - M., 1995. - 160 р.
Authors' Information
Vladimir Talalaev Assoc. Prof. of Software Engineering Department,
National Aviation University, Kyiv, Ukraine. E-mail:
VATalalaev@gmail.com)
Major Fields of Scientific Research: Model-Driven Architecture,
Software architecture, Software development, Software risks, business
process modeling
Larysa Postavna assistant of Software Engineering Department,
National Aviation University, Kyiv, Ukraine. E-mail:
larysa.postavna@npp.nau.edu.ua (postavnalarisa@gmail.com)
International Conference on Software Engineering
AUTOMATIC ONTOLOGY CREATION FROM NATURAL LANGUAGE TEXT
Anna Litvin, Vitalii Velychko, Vladislav Kaverinsky
Abstract: An approach and its software implementation was created to build an OWL
ontology using as an input a natural language text or set of texts. The approach is
primarily aimed but not limited to inflected languages. The method for the moment is
acceptable for tagged or regularly structured texts. The created ontologies are tested as
valid using Protégé and RDFlib and are useful for dialog and reference systems.
Keywords: NLP, NLU, ontology, automatic ontology creation, OWL.
ITHEA Keywords: H.3.4 Systems and Software.
Conference and topic: Applied Aspects of Software Engineering: Software Designing.
Introduction
Manually database creation and fulfillment could be a labor-intensive and rather boring
process. Thus its automation is highly desirable. An information source for this purpose
could be a natural language text that contains some information. Considered here
database type is a graph database that is useful for ontologies data structures that
represent entities of some subject area and relationships between them. One of the
popular languages for ontologies construction is OWL [Antoniou, 2016], which was used
in this work. An OWL ontology can be easily converted to an RDF-triples set for which
SPARQL queries can be used for information obtaining.
Some methods for ontologies creation automation have been proposed, for instance, in
works [Young, 2018; Elnagar, 2020; Balakrishna, 2008]. But the method and the goal of
a certain approach are highly dependent on the destination of the ontology and its usage,
which affect the structure of the ontology. Thus the task of new systems building for
ontologies automatic creation aimed at the certain case to construct the most useful and
suitable in structure and content graph database seems to be actual and important.
International Conference on Software Engineering
22
The considered ontology structure
In the proposed approach all the entities in the ontology are devised in two main groups
classes and properties. The relationship types are limited to the following: “Subclass”
links more certain entity to more abstract, “Subproperty” - links more certain property to
more abstract one, “Domain” links to a property an entity (or set of entities) which has
the property, “Range” - links to a property an entity (or set of entities) which is the value
of the property. In this way two entities (or sets of entities) can be linked with a property
where one entity is its “Domain” and the other is “Range”. So “Domain/Range” linking
could be used for a horizontal not hierarchical binding of entities. These relationship
types are standard for OWL and are rather enough for most cases.
In the proposed ontology structure the main abstract predefined classes are: “Action”,
“Adjective”, “Adverb”, “Number”, “Object” (nouns and name groups), “Marker” (question
and negotiation words), “ReadyAnswerLink” (links to large text blocks and media
information), “Count” (denoting the number of possible occurrence multiplicity of a term in
the text), “Quantity” (the actual multiplicity of occurrence of a term in the text or
considered part of the text), “ConditionIntersection” (combinations or intersections of
concepts from a potential user phrase that determines a certain response of the system).
The predefined main properties in the considered ontology structure are following:
“MainProperty” its descendent's link combinations of concepts to descendants of the
“ConditionIntersection” class; “BindQuantityToInteraction” - binds reference to text blocks
- descendants of the class “ReadyAnswerLinkto the multiplicity of occurrence in their
text of certain concepts descendants of the class “Quantity”; “IntersectionCount” - binds
the actual multiplicity of occurrence of the concept in the text block descendants of the
class “Quantity” to indicate the number of the multiplicity of occurrence of the concept in
the certain text block descendants of the class “Count”.
Subclasses of the “Object” class can have their descendants, which are name groups.
The hierarchy of descendants of the “Object” class is as follows: the highest are the
individual nouns that are the main in name groups and define the most general variant of
International Conference on Software Engineering
the concept then the hierarchy is followed by name groups as words grow in them, which
define more specific entities. The ontology structure could be supplemented with the
other concepts specific to the considered subject area, for example: “Date”, “Person”,
“Location”, etc. These ones can also have predefined descendants, for instance: “Author
a descendant of “Person”, “CreationDate” a descendant of “Date”, etc. The
predefined ontology entities and structures form the framework basing on which further
automatic graph creation becomes possible and moreover the methodology of the
following usage of the ontology becomes more clear, because the basic structure of the
SPARQL queries for information obtaining becomes more predictable.
The ontology creation technique
It is assumed that the input text is tagged or regularly structured, which makes it easier to
obtain a good quality of the resulting ontology. A tagged text is actually a set of texts
each of them has at least a title and main parts. Such set of texts could be obtained for
example by web pages scraping an informational resource. On each page there could be
found a title of the article, its text and may be some other information in certain places:
date, author, location, subject area, article type etc. An example of regularly structured
text could be a collection of letters of or to some person. For each of the letters, there
could be found in the same or approximately the same parts such information tokens like
“destination address”, “sender address”, “dates of sending or/and receiving”, “sender’s
name”, “addressee”, “the main text” etc. Additional semantic analysis of the main text
block and the title (if exist) could be used to estimate the topic, type and subject of it.
The main text blocks which could also contain images and media content are stored in a
separate document base. In OWL ontology only links on them are to be present. To this
links are bind directly obtained information blocks (such as dates, locations, topics). And
they become ontology classes of a certain predefined ancestor. If the information is
repeated only one point is needed to be stored in the ontology. So multiplied links are
created. To create such links specific properties are used. For example, there is a
sending date and a letter text. A property “SendingDates” is to be created. It has
International Conference on Software Engineering
24
decedents each of them has a certain date as its “Domain” field value and list of letters
sent on this date.
In other way is performed linking a title to the main information block. The titles are not
stored in the ontology. They are tokenized to separate words and name groups, which
are stored in the ontology according to their role in the sentence and without repeating.
The links to main texts are bind to the entities sets (intersection) according to the
meaningful terms from the corresponding title.
If a concept (only for nouns and noun groups) of those included in the intersections of
concepts associated with a particular text, is included in the text more than once, it is
considered that this concept is more important and decisive. Therefore, the
corresponding subclasses of “Count” and “Quantity” are created for them. The same
subclass “Count” can be associated with different subclasses of Quantity, which avoids
duplication for example, three times different concepts may occur in different texts. In
turn, a subclass of the “Quantity” class can be linked to multiple texts links the same
concept can be included, for example, twice in different texts. This information is useful to
range the ontology responses for queries by their relevancy and informative value.
Conclusion
Such automatically created ontologies could be useful for dialog and reference systems.
For them, it is rather easy to create a limited number of unified formal query templates,
input information for which can be concepts obtained from the user’s phrase in such
manner as they obtained from titles and texts during ontology creation. Such system
become rather reliable: even in case of lack of some concepts in the user’s phrase or
their super saturation the rather relevant response could be obtained.
Some examples of software that implement the proposed approach was created for
Ukrainian, English and Norwegian languages. As information sources for different
ontologies were used: finance consulting information (Norwegian), climate change
questions (English), articles by literary critics about the work of Oles Honchar (Ukrainian),
International Conference on Software Engineering
correspondence of Oles Gonchar (Ukrainian), pedagogical works of Vasily Sukhomlinsky
(Ukrainian), medical rehabilitation glossary (Ukrainian). Testing of the dialog systems
Bibliography
1. [Antoniou, 2016] Antoniou G. Semantic Web. Moscow: DMK-Press, 2016.
2. [Young, 2018] B. Young, A. JungHyen. Methodology for Automatic Ontology
Generation Using Database Schema Information. Mobile Information Systems, 2018.
pp. 1 13.
3. [Elnagar, 2020] S. Elnagar, V. Yoon, M. A. Thomas. An Automatic Ontology
Generation Framework with An Organizational Perspective. Proceedings of the 53rd
Hawaii International Conference on System Sciences. 2020 pp. 4860 4869.
4. [Balakrishna, 2008] M. Balakrishna, M. Srikanth. Automatic Ontology Creation from
Text for National Intelligence Priorities Framework (NIPF). Proceedings of 3rd
International Ontology for the Intelligence Community (OIC) Conference. 2008. pp. 8
12.
Authors' Information
Anna Litvin Glushkov Institute of Cybernetics of NAS of Ukraine;
postgraduate student. 40 Glushkova ave., Kyiv, Ukraine, 03187; e-mail:
litvin_any@ukr.net
Major Fields of Scientific Research: natural language processing for
analysis
Vitalii Velychko Glushkov Institute of Cybernetics of NAS of Ukraine;
Ph.D. Senior Researcher. 40 Glushkova ave., Kyiv, Ukraine, 03187; e-
mail: aduisukr@gmail.com
Major Fields of Scientific Research: information systems with the
processing of natural language objects
Vladislav Kaverinsky Institute for Problems of Materials Science, Ph.D.
Senior Researcher, 3 Krzhizhanovsky st., Kyiv, Ukraine; e-mail:
insamhlaithe@gmail.com
Major Fields of Scientific Research: mathematical modeling
International Conference on Software Engineering
26
APPLICATION FEATURES OF AGILE APPROACHES IN SOFTWARE ENGINEERING
Iryna Gruzdo, Ihor Michurin
Abstract: The features of agile approaches in software engineering and points of control
of the schedule in the development of software products are highlighted.
Keywords: Software Engineering, AGILE, SCRUM, development, business.
Introduction
The use of agile approaches in software development has had a significant impact on the
development of the IT business. therefore, highlighting the features of the agile
methodology and how it differs from traditional development approaches is important for
effective software development. It is also necessary to highlight the features of the
control of the project schedule when using agile methodologies, which is usually
performed by the project manager. The correct application of agile methodologies allows
you to achieve high efficiency in software product development. The purpose of the
article is to systematize the main distinctive features of the agile methodology and the
features of planning the project development schedule when using them.
General features of agile approach
The emergence and development of Agile approaches is associated with the need to
respond to the rapid changes in customer requirements. It is important to note that
effective implementation of these approaches requires a high level of interaction between
the development team and the customer. Also, each iteration in the Agile method
approach takes 2-4 weeks, that is, it represents a very fast implementation of iterative
and incremental methods. Moreover, each iteration is usually fixed in terms of time and
total development costs. During each iteration of the project development, the team can
perform several tasks simultaneously. For example, during the initial iterations, the
team's focus is primarily on planning and analysis tasks. The entire content of the project
is divided into a list of requirements. The request log contains a list of outstanding work,
International Conference on Software Engineering
that is, work that has not yet been completed. At the beginning of each iteration, the
project development team determines from the requirements log those tasks that will be
completed during the iteration and the implementation of which is currently the most
important for the project. At the end of the iteration, the tasks should be completed and, if
necessary, shown to the customer. Typically, customers or customer representatives
must participate in the project development process at all times to interact with the
development team and provide feedback. In turn, the presence of feedback allows the
team to properly meet the needs of the customer. This allows you to quickly respond to
changing customer needs, which can be useful for both the development team and the
customer. During each iteration, when applying agile approaches, it is important to
control the schedule of tasks performed by the team. To do this, the project manager
needs to determine the current size of the completed work with the planned amount of
work. At the end of each iteration, it is necessary to perform a retrospective analysis in
order to analyze the completed iteration and possible improvement of the software
product development processes. In addition, the control of the schedule is associated
with the analysis of the priorities of the work plan that have not yet been completed, and
their change, if necessary. Determining the level of productivity of the software product
development team is important for effectively scheduling tasks. It is necessary to
determine the effective duration of one iteration of the project development process,
which can last from two to four weeks. In the event of changes, the task of determining
their level and impact on the pace of project development becomes urgent [Hamed and
Abushama, 2013], [Yashyna et al, 2017].
Specifics of using agile methodologies in companies of different sizes
Today Agile methodology is used not only in IT but also in classic business. The most
popular methodologies included in Agile is: Scrum, SCRUM of SCRUMs, Scrum@Scale,
Disciplined Agile Delivery (DAD), Scaled Agile Framework (SAFe), Large Scale Scrum
(LeSS), Agile Unified Process (AUP), Crystal Clear Method, Dynamic Systems
Development Method (DSDM), Feature-Driven Development (FDD), Scaled Agile
Framework (SAFe), etc. One of the popular Agile methods is Kanban. Its simplicity
International Conference on Software Engineering
28
allows it to be used effectively, especially in small teams. In addition, there is no need to
use additional staff to monitor the implementation of this methodology. But in medium
and large teams, its use is ineffective due to the large number of tasks that the team
performs. Another disadvantage of Kanban is the lack of a backlog. One of the most
popular Agile methods is Scrum. It consists of iterations and allows to develop software
efficiently and systematically. The backlog allows to see all the tasks and control the
development process. Scrum is primarily based on self-organization and team skills. In
general, Scrum can be used for teams of all sizes. In the case of a large project, a Scrum
of Scrums can be applied. On the other hand, misusing Scrum or using Scrum on an
inexperienced team can lead to failure. A Scrum Master is also often needed to
supervise the correct application of the Scrum methodology. Scaled Agile Framework.
The Scaled Agile Framework is an agile methodology that is primarily targeted at large
companies with many teams. This methodology allows to organize an effective
development process in large companies, but in medium and small companies, its
application can be ineffective. During the study, the Scrum methodologies that are used
in the development of software were studied, it can be concluded that the choice of a
specific methodology depends on the project itself, its size and time required for its
implementation, finances, as well as the quality of the team. But it should be noted that
all Scrum technologies allow you to achieve high-quality results in a short time and get
finished software. It can be concluded that when companies are faced with large and
complex projects with constantly changing requirements or goals, they turn to more
flexible approaches.
Conclusion
Agile methodology is most effective when implemented correctly and can increase the
productivity of the software development team, accelerate customer satisfaction and
reduce costs. Scrum Master Trends report [2019 Scrum Master Trends Report] from the
year 2019 shows that 81% of the respondents claimed to use Kanban in addition to
Scrum. Kanban, Scrum, Scrum of Scrums and the Scaled Agile Framework are effective
International Conference on Software Engineering
agile methodologies for software development in IT companies. These methodologies
are the most used in software development practice.
Bibliography
1. [Hamed and Abushama, 2013] Hamed, A. M. M., Abushama, H., Popular agile
approaches in software development: Review and analysis. International conference
on computing, electrical and electronic engineering (ICCEEE), Khartoum, Sudan, 26-
28 Aug. 2013, DOI: https://doi.org/10.1109/icceee.2013.6633925
2. [Yashyna et al, 2017] Yashyna, K. V., Yalova, K. M., Sugal, E. O., Software
development agile methods overview. Collection of scholarly papers of Dniprovsk
State Technical University (Technical Sciences), Vol. 1 No 30., 2017. Pp 153-156.
http://sj.dstu.dp.ua/article/view/145898
3. [2019 Scrum Master Trends Report] https://www.scrum.org/resources/2019-scrum-
master-trends-report
Authors' Information
Iryna Gruzdo Assoc. Prof. of the Department of Software Engineering,
Kharkiv National University of Radio Electronics, Kharkiv, Ukraine. E-mail:
irina.gruzdo@nure.ua.
Major Fields of Scientific Research: Information Technology;
Plagiarism; Software Project Management; Business Intelligence
Ihor Michurin student of the Department of Software Engineering,
Kharkiv National University of Radio Electronics, Kharkiv, Ukraine. E-mail:
ihor.michurin@nure.ua
Major Fields of Scientific Research: Information Technology; Software
Engineering; Artificial Intelligence
International Conference on Software Engineering
30
AN APPROACH OF DESIGNING RESTFULL API MODEL FOR DISTRIBUTED WEB
APPLICATIONS
Yurii Milovidov
Abstract: Paper proposes an extensibility of RESRFULL API approach to perform more
reliable web sites and steps towards web development.
Keywords: Model-View Controller, REST, HTTP requests.
Introduction
REST approach is widespread used in modern web-development activities. Summarizing
practices if web-development companies the sense of this approach is the next: sorting
of Controller methods by HTTP requests for more quick understanding of Controller
structure [Milovidov Yu., 2019]. Thus, such an approach is aimed to perform cognitive
and communicative functions in web-development [Chebanyuk & Markov, 2015].
Contribution of this thesis: Extended approach considering cognitive and communicative
functions of development team is proposed. Advantages of this approach. Proposed
model can be basic for implementing practical approaches for development of logical and
automated analysis of web-page lifecycle with open extensible mode to add for example
security functions to web-page lifecycle.
Proposed Approach
It is proposed to give universal manes of different Controller method analyzing whole live
cycle of web-page.
Summarizing table is represented below
Init stage
Request type
GET
Function Init(dataset)
To obtain initial data to download
View Model
Generated markup and
initial sets of data and
International Conference on Software Engineering
metadata for starting
operations
GET
Add listener1
listenern
To set up basic functionality
(event processing) of View Model
components
To create interactive
instances of View Model
User Interface
components
GET
Error processing
functions
To provide secure and stable
working mode for web-page
Add security functions to
input fields and
optional
GET
refine
To refine MarkUp
To add extra settings to
*.html page
Working stage
Post
Renew parts of *.html
page
Provide interactive answers to
User requests
Send data from html form
Delete
Delete some
resources
DeleteResourceNAME
Delete some resource
Put
UpdateResourceName
Renew of some resource
Conclusion
Proposed table summarizes stages of web-page lifecycle. Recommended names of
methods can be base to analyze Controller web-pages by means of static analysis tools
International Conference on Software Engineering
32
and verify whether all necessary patterns of web-page lifecycle are realized in web-
application.
Bibliography
1. [Chebanyuk E. & Markov Kr. 2015] Chebanyuk E. & Markov Kr. Software model
cognitive value. International Journal “Information Theories and Applications”, Vol.
22, Number 4, 2015, 338-355
2. [Milovidov Yu. 2019] Yurii Milovidov. Comparison of web services creation
technologies using testing application SoapUI. International Journal "Information
Content and Processing", Vol. 6, Number 2, 2019, 60-67
Authors' Information
Yurii Milovidov, National University of Life and Enironmental Sciences of
Ukraine, Kyiv, scientific Department of Programming Technologies.
Senior Lecturer. Email: milovidov.email.ua
Major Fields of Scientific Research: Programming technologies, web-
design, Internet technologies
International Conference on Software Engineering
SECTION: APPLIED ASPECTS OF SOFTWARE ENGINEERING
THE METHOD MULTICRITERIAL EVALUATION OF VERIFY THE QUALITY
PROGRAM SYSTEMS IN USE
Yuliia Bezkorovaina
Abstract: The main purpose of verifying quality is to ensure that the software systems
are meeting the requirements and implied needs of the users. The verification of quality
includes defining purpose and measures, their evaluation, and decide on it. One of the
tasks is what method of evaluation to select, and develop automatic calculations. This
thesis offers a solution to the problem on the example of the quality model of software
systems in use and method multicriterial evaluation.
Keywords: verification, software, quality in use, multicriterial evaluation, quality model,
SQuaRE.
ITHEA Keywords: D.2.1 Requirements/Specifications, D.2.4 Software/Program
Verification, D.2.8 Metrics
Conference and topic: International Conference on Software Engineering “SoftEngine -
2021”, Approaches to Software Development Life Cycle Processes Improvement,
http://www.ithea.org/softengine/
Introduction
The software systems pass through the process of the life cycle. The standard
[ISO12207, 2017] described phases of the development. The software life cycle is
allocated to the process: agreement, organizational project-enabling, technical
management, and technical. To initiate the process, supplier and customer must agree
on a purpose, outcomes, and activities and tasks [ISO12207, 2017], [IEEE1012, 2017].
This thesis will be about the verification of quality and acceptance of software systems.
There are a series of standards about quality verification the software systems. The part
of those standards describes different quality models [ISO25010, 2011]. There is a
quality in use model [ISO25022, 2016], quality product model, quality data model, quality
International Conference on Software Engineering
34
services model, and so on. This thesis is dedicated to verifying quality in the use of the
final software system.
The purpose of evaluating the quality of the final software product is being to decide on
the acceptance of the product and determine the time of the product release [ISO25040,
2011]. To assess the quality of the finished software system during use, you need to
determine the requirements for it. The requirements specification considers the budget,
target dates, and the purpose of the assessment. The specification can change or
improve during the development of software systems [ISO25040, 2011].
Related work
There is an empirical analysis of quality standards concerning usability characteristics
and another software system quality characteristics [Bevan et al, 2016], the
implementation of a product quality model according to the standard [Kazuhiro, 2013],
[Yusupova et al, 2016]. Other sources are considered [Danyk et al, 2016], [Pysarchuk
and Pinchuk, 2006], [Pysarchuk et al, 2019], which prove in practice the effectiveness of
the proposed method.
Task and challenges
The complexity of the quality evaluation task is justified by the variety of stakeholders
and users, areas of application of quality models, the complexity of software systems and
the variety of their life cycle, quality management tools, etc. The main task of research is
to prove the effectiveness of the multicriteria method for the quality model in use
evaluation of the software systems.
Body
The quality model in use is composed of the standard [ISO25022, 2016] and evaluation
of the model [ISO25040, 2011]. Quality in use model composed of five characteristics
that relate to the outcome of interaction when a product is used in a particular context of
use. So, the quality in use model (characteristics, subcharacteristics, and measures)
represent in table 1. Each measure required the calculation of metrics. The value of each
International Conference on Software Engineering
criterion affects the software system quality in use. So, we can say that this is a problem
with many diverse criteria.
Table 1. Model of quality in use
Characteristics
Subcharacteristics
Measures
Effectiveness
-
Tasks completed
Objectives achieved
Errors in a task
Tasks with errors
Task error intensity
Efficiency
-
Task time
Time efficiency
Cost-effectiveness
Productive time ratio
Unnecessary actions
Fatigue
Satisfaction
Usefulness
Overall satisfaction
Satisfaction with features
Discretionary usage
Feature utilization
The proportion of users complaining
The proportion of user complaints about a
particular feature
Trust
User trust
Pleasure
User pleasure
International Conference on Software Engineering
36
Characteristics
Subcharacteristics
Measures
Comfort
Physical comfort
Freedom from
risk
Economic risk
mitigation
Return on investment
Time to achieve a return on investment
Business performance
Benefits of IT Investment
Service to customers
Website visitors converted to customers
Revenue from each customer
Errors with economic consequences
Health and safety
risk mitigation
User health reporting frequency
User health and safety impact
Safety of people affected by the use of the system
Environmental
risk mitigation
Environmental impact
Context
coverage
Context
completeness
Context completeness
Flexibility
Flexible context of the use
Product flexibility
Proficiency independence
As we can see, there is a problem to decide on the readiness of the software system.
The standard [ISO25040, 2011] proposes different methods of evaluation quality. But in
this thesis, the author proposes to evaluate the final version of the software system with
the method of multicriteria evaluation using nested convolutions according to a nonlinear
scheme of compromises [Voronin, 2017], [Kharchenko and Pysarchuk, 2014].
International Conference on Software Engineering
The quality model transforms into an infographic model of factors, indicators, and criteria
without subcharacteristics that are not required by the standard [ISO25010, 2011]. By
heuristic analysis of the quality model in use formed an infographic model of factors,
indicators, and criteria of the quality in use (Table 2).
Table 2. Infographic model factors, indicators, and criteria
Factors
Factors in the group
Indicator
Criteria
1
Effectiveness
1
Tasks completed
max
2
Objectives achieved
max
3
Errors in a task
min
4
Tasks with errors
min
5
Task error intensity
min
2
Efficiency
6
Task time
min
7
Time efficiency
min
8
Cost-effectiveness
min
9
Productive time ratio
max
10
Unnecessary actions
min
11
Fatigue
min
3
Satisfaction
12
Overall satisfaction
max
13
Satisfaction with features
max
14
Discretionary usage
→max
International Conference on Software Engineering
38
Factors
Factors in the group
Indicator
Criteria
15
Feature utilization
→max
16
The proportion of users complaining
→min
17
The proportion of user complaints about
a particular feature
→min
18
User trust
→max
19
User pleasure
→max
20
Physical comfort
→max
4
Freedom from
risk
21
Return on investment
→max
22
Time to achieve a return on investment
→min
23
Business performance
→max
24
Benefits of IT Investment
→max
25
Service to customers
→max
26
Website visitors converted to customers
→max
27
Revenue from each customer
→max
28
Errors with economic consequences
→min
29
User health reporting frequency
→max
30
User health and safety impact
→max
31
Safety of people affected by the use of
the system
→max
International Conference on Software Engineering
Factors
Factors in the group
Indicator
Criteria
32
Environmental impact
→max
5
Context
coverage
33
Context completeness
→max
34
Flexible context of the use
→max
35
Product flexibility
→max
36
Proficiency independence
→max
The formation of the decision-making model consists of aggregating the list of
contradictory partial criteria to the generalized assessment. It was therefore selected for
convolution nonlinear compromise scheme [Voronin, 2017], [Kharchenko and Pysarchuk,
2014].
The main idea of the approach is in the following sequence:
1. Forming a system of partial indicators of quality;
2. Determining the values of partial indicators of decision quality;
3. Formation of a generalized quality indicator from a system of partial and bringing it
to a relative value.
4. Analysis of evaluation and decision-making on the compliance of the software
system in use.
Interpretation of the received decision consists of the resulted values of the integrated
estimation to a uniform scale of change - from 0 to 1. The received numerical estimation
can be reduced to a linguistic category according to a fundamental scale of estimation
(Table 3).
Table 3. Fundamental rating scale
Integrated quality assessment
Linguistic category of quality
1.0-0.7
High
International Conference on Software Engineering
40
0.7-0.5
Good
0.5-0.4
Satisfactory
0.4-0.2
Low
0.2 and less
Unsatisfactory
So, after calculation, the model can help us decide about the software systems quality in
use. If we had received a value of 0.6, it means that the software system has a good
quality. That quality depends on a set of measurements. Supplier and customer may
decide or to accept as a final version, or develop onwards a software system.
Conclusion
This thesis considered the quality model in use by the standard [ISO25010, 2011],
[ISO25022, 2016], [ISO25040, 2011]; method of evaluation of the quality in use [Danyk et
al, 2016], [Pysarchuk and Pinchuk, 2006], [Pysarchuk et al, 2019]. The next steps in
research are to evaluate the quality product model by standard ISO/IEC 25023
“Measurement of system and software product quality”, compare the effectiveness of
evaluation methods and identify the most effective.
Bibliography
1. [ISO12207, 2017] ISO/IEC/IEEE 12207:2017. Systems and software engineering
Software life cycle processes. ISO, Switzerland, Geneva, 2017, 145 p.
2. [IEEE1012, 2017] IEEE Std 1012:2017. Standard for System, Software, and
Hardware Verification and Validation, in IEEE Std 1012-2016 (Revision of IEEE Std
1012-2012/ Incorporates IEEE Std 1012-2016/Cor1-2017) - Redline, 2017, 465 p.
3. [ISO25010, 2011] ISO/IEC 25010:2011. Systems and software engineering. Systems
and software Quality Requirements and Evaluation (SQuaRE). System and software
quality models. ISO, Switzerland, Geneva, 2011, 34 p.
4. [ISO25022, 2016] ISO/IEC 25022:2016. Systems and software engineering
Systems and software Quality Requirements and Evaluation (SQuaRE)
Measurement of quality in use. ISO, Switzerland, Geneva, 2016, 41 p.
5. [ISO25040, 2011] ISO/IEC 25040:2011. Systems and software engineering
Systems and software Quality Requirements and Evaluation (SQuaRE)
Evaluation process. ISO, Switzerland, Geneva, 2011, 45 p.
International Conference on Software Engineering
6. [Bevan et al, 2016] Bevan N., Carter J., Earthy J., Geis T., Harker S.. New ISO
Standards for Usability, Usability Reports and Usability Measures. Conference:
International Conference on Human-Computer Interaction, Theory, Design,
Development and Practice. Vol. 9731, pp 268-278. DOI: 10.1007/978-3-319-39510-
4_25
7. [Yusupova et al, 2016] Yusupova N., Gvozdev V., Yanchek K., Smetatina O.,
Morozov A. MODEL'' Kachestva SQUARE programmnogo producta. INFORMATION
TECHNOLOGIES FOR INTELLIGENT DECISION MAKING SUPPORT
(ITIDS'2016). Ufa, 1719 maja, 2016, pp. 54-59
8. [Kazuhiro, 2013] Kazuhiro Esaki. Verification of Quality Requirement Method Based
on the SQuaRE System Quality Model. American Journal of Operations Research,
2013, 3, 70-79. DOI: 10.4236/ajor.2013.31006
9. [Danyk et al, 2016] Danyk Yu., Pysarchuk O., Lagodniy O., Vyporhoniuk O.
Matemetichna model bagatokriterialnogo otsiniuvania efektyvnosti internet-saitiv
tsilovogo spriamuvania. Ingeneria programnogo zabezpechenia. Visnik, Issue 1, Vol.
76. 2016, pp. 114-120
10. [Pysarchuk et al, 2019] Pysarchuk O., Bezkorovaina Yu., Dyshlevyi O., Skalova V.
Metodyka bahatokryterialnogo ocinyuvannya vidpovidnosti prohramnoho
zabezpechennya vymoham zamovnyka. Naukoyemni texnolohiyi. 2019. №1 (41).
Kyiv: NAU, pp. 3-9.
11. [Pysarchuk and Pinchuk, 2006] Pysarchuk O., Pinchuk O. Metodyka vyboru zasobiv
z vikorystanniam bagatokriterialnyh modeley. Visnik: tehnichni nauky. Vol. 4, Issue
39, 2006, pp. 152-158.
12. [Voronin, 2017] Voronin A. Multi-Criteria Decision Making for the Management of
Complex Systems, Multi-Criteria Decision Making for the Management of Complex
Systems, March, 2017, 201 р. DOI: 10.4018/978-1-5225-2509-7
13. [Kharchenko and Pysarchuk, 2014] Kharchenko V., Pysarchuk O. Neliniyne ta
bagatokriterialne modeliuvannia protcesiv u systemakh keruvania rukhom, NAU,
Kyiv. 2014, 372 p.
Authors' Information
Yuliia Bezkorovaina Senior Lecturer of Software Engineering
Department, National Aviation University, Kyiv, Ukraine; e-mail:
yuliia.bezkorovaina@nau.edu.ua
Major Fields of Scientific Research: Quality, software verification,
software engineering
International Conference on Software Engineering
42
SOFTWARE TECHNOLOGY OF FLOOD SIGNALING AND FORECASTING IN THE
TRANSCARPATHIAN REGION
Volodymyr Polishchuk
Abstracts: The research constructs a generalized algorithm for signaling and forecasting
floods in the Transcarpathian region, based on a hybrid intelligent model, using the
analysis of reasoning and experience of water management experts. Based on the
studied concepts, the construction of software technology in the form of web technology
for an expert administrator and a cross-platform application for the user is envisaged.
Keywords: Forecasting, hybrid data, software, crisis management.
Introduction
In the life of modern society, the problems associated with overcoming the
consequences of emergencies, which lead to huge material losses and sometimes
human losses, play a significant role. A special place among many emergencies belongs
to floods of man-made and natural origin, catastrophic flooding, and flooding of
territories, which are some main socio-environmental problems of our time [Kelemen M.,
2019]. Floods have long been a problem for human society, but today their size and
frequency are growing rapidly, causing material damage and, worst of all, human
casualties. The Transcarpathian region is no exception. The negative effects of flooding
cannot be completely avoided, but their impact can be reduced by forecasting and
signaling the population while giving time to prepare or mobilize society and relevant
bodies.
In the context of crisis management and security of citizens, there is a complex and
urgent task of developing a mobile software product that allows, according to the
geolocation data of the smartphone, to warn the user about the level of flood risk.
Thus, the aim of the study is to develop a generalized algorithm and software technology
for signaling and forecasting floods in the Transcarpathian region, based on a hybrid
intelligent model, using the analysis of reasoning and the experience of water experts.
International Conference on Software Engineering
Generalized algorithm of flood signalling and forecasting
Suppose we have n measuring stations, which every period of time ti send data on the
actual water level in ni. To analyze the measurement data, we use the method of
regression analysis, namely the method of pairwise linear regression. We present a
general algorithm for signaling and forecasting floods based on a hybrid intelligent model
[Polishchuk V., 2019] to determine the level of flood risk. The model is able to derive a
normalized assessment of water level rise, uses the analysis of reasoning and
experience of water management experts, reveals the vagueness of input estimates,
increases the degree of validity of further management decisions based on the results.
The output of the model is a normalized assessment of the danger of a flood situation
and a linguistic interpretation of the level of danger of a flood situation.
Step 1. Obtaining data and building a time series
In the first step, we construct a time series where yi is the water level, xi is the fixation
time.
Step 2. Determination of regression coefficients
In the second step, the least-squares method (LSM) is used to determine the regression
parameters. Find the values of the regression coefficients. LSM allows obtaining such
estimates of parameters in which the sum of the squares of the deviations of the actual
values of the resultant feature y, from the theoretical values of y at x is the minimum.
Step 3. Obtaining predicted values
In the third step, we construct a pairwise linear regression equation to predict the water
level. To do this, we substitute future periods x into the equation to obtain the predicted
values.
Step 4. Fuzzyfication of input data
In the fourth step, we will fuzzification the input data, based on the predicted data, the
maximum and minimum allowable water level in the area (stations) pit(max), and pit(min),
according to the formula:
International Conference on Software Engineering
44
(1)
Where we set the safe (lowest) level, respectively -
dangerous (highest) level. The concept of constructing membership functions assumes
the following meaning: the higher the water level, the greater the risk of flooding, and the
value of the membership function goes to one [Kelemen M. et al., 2020].
Step 5. Estimation of water level projection
In the fifth step, we estimate the water level projection for the relevant area over a period
of time:
(2)
The obtained value characterizes the estimation of the water level projection for the
respective area in a certain period of time [Polishchuk V., 2019].
Step 6. Obtaining a flood risk assessment
The sixth step requires the water expert to assess the risk of flooding and select one of
the proposed linguistic variables: H {Low risk of flooding}; C {Medium risk of flooding}; B
{High risk of flooding}. After that, we calculate the normalized assessment of the risk of
flood situation by the formula:
(3)
Where k is the degree of risk of flooding. For example, put: for low risk of flooding k = 1/3,
for medium risk k = 1, and for high risk of flooding k = 3/5.
According to the received normalized estimation, we will present a linguistic interpretation
of the level of danger of a flood situation: LS: low level of flood
International Conference on Software Engineering
danger; AS: average level of flood danger;
HS: high level of flood danger.
The range of linguistic interpretation, to allow for adjustment, is specified in the settings.
Step 7. Decision-making on flood signaling in Transcarpathian region
In the seventh step, based on the linguistic interpretation, the decision-maker (water
management expert) can signal and/or inform users about the dangerous situation,
respectively, through the mobile application.
Conclusion
As a result of the implementation of smart technologies of crisis management,
strengthening of protection against harmful effects of water of settlements, minimization
of the caused losses, and creation of safe living conditions of communities of the region
is carried out.
An algorithm based on a hybrid intelligent model has been developed, which uses the
analysis of reasoning and experience of water management experts, reveals the
vagueness of input estimates, increases the degree of validity of further management
decisions based on the results. Web technology and a cross-platform user application
will be designed for the expert administrator part. With the help of the application, the
user, based on geolocation data, receives notifications about the possible level of
danger. In addition, it is assumed that the user can view the state of the flood situation in
the location of his stay, the water level, and the predicted danger. Real data will be
obtained on the basis of publicly available AIS "Tisa", consisting of 50 stations
throughout the Transcarpathian region and neighboring EU countries.
Bibliography
1. [Kelemen M., 2019] Kelemen M. (2019). Safety and Knowledge Alliance of Aviation
Education: Human factors in Aviation Safety and Air Law. AMELIA Aneta Siewiorek,
Poland, 2019.
International Conference on Software Engineering
46
2. [Polishchuk V., 2019] Polishchuk V. (2019). Technology to Improve the Safety of
Choosing Alternatives by Groups of Goals. Journal of Automation and Information
Sciences, 51 (9), 66-76. DOI: 10.1615/JAutomatInfScien.v51.i9.60
3. [Kelemen M. et al., 2020] Kelemen, M., Polishchuk, V., Gavurová, B., Andoga, R., &
Matisková, D. (2020). The Expert Model for Safety Risks Assessment of Aviation
Environmental Projects' Implementation Within the Investment Phase of the Project.
IREASE, Vol 13, No 6, 198-207. doi: 10.15866/irease.v13i6.18268).
Authors' Information
Volodymyr Polishchuk Assoc. Prof. of Department Software
Systems, Uzhhorod National University, Uzhhorod, Ukraine. E-mail:
volodymyr.polishchuk@uzhnu.edu.ua
Major Fields of Scientific Research: Intelligent Information
Processing, Data Mining, Expert Systems, Fuzzy set, Knowledge
Engineering and management, Intelligent software and
Computational Intelligence
International Conference on Software Engineering
SECTION: SOFTWARE ENGINEERING APPLICATION DOMAINS
AUTOMATIC SEARCH FOR INVOLVANTS OF INFORMATION MESSAGES IN THE
VOICE DATABASES
Serhii Zybin, Yana Bielozorova
Abstract: Parameters of individual voice characteristics in modern automatic systems of
speaker identification was researched. Efficiency of such systems mainly depends on the
methods for determination of these parameters and stability of spectral characteristics
and the main tone frequency. The structure and concept for an automated search in
voice databases was described.
Keywords: Speaker identification system, wavelet, coding and information theory,
pattern analysis, phonogram.
Conference and topics: SoftEngine 2021; Software Engineering Application Domains -
Image Processing and Computer Vision.
Objectives
Spectrum analysis of audio data based on Fourier transform mathematical apparatus is
the key factor in majority of modern systems for solving the tasks of speaker identification
by means of voice characteristics. It is caused by several factors, known
neurophysiological rules of audio information processing by primary auditory receptors
and absence of more effective analysis methods and, to a certain extent, by historical
traditions in this field.
However, in spite of a lot of investigations in this field and creation of rather effective
systems for voice characteristics identification and development of systems of
identification and text entering of voice information, there is not enough clearness in the
principal theoretical and practical issues concerning speech technology till now. Mainly, it
is caused by absence of effective mathematical tool for analysis of voice audio
information [1,2].
International Conference on Software Engineering
48
The goal is to study the effectiveness and develop a software system for the formation of
voice databases of participants in information messages, as well as automatic search in
the voice database.
Model of automated search in voice databases
Concept implemented in the experimental system of instrumental identification of
speaker voice characteristics was proposed. This concept is based on presentation of
speech fragments as a set of multifractal structures. In order to determine parameters of
multifractal structures the wavelet analysis with special basis in form of two-parameter
Morlet wavelet is used [3,4].
Till now, actually, the single methodological direction for building the mathematical
models for calculation of phoneme parameters and voice characteristics is the
application of some of modifications of spectrum analysis based on Fourier transform.
However, the application of this spectrum analysis method has number of serious
disadvantages (known in the years since Helmholtz). Signal spectrum of “instrumental”
transformers of acoustical apparatus hair cells can be considerably different from
Fourier spectrum, including (that is rather important) arrangement of formants in
frequency domain. Human acoustical apparatus is able to distinguish frequencies with a
step smaller than 1 Hz (within the range up to 500 Hz).
It is clear, that physical and mathematical models of voice information transformation
should not necessarily strictly follow neurophysiologic rules. However, analysis of voice
information based on orthogonal Fourier transformation for many decades at the
phonemic level shows rather high variability of spectra for the same phonemic structures
in short time intervals. It is a significant obstacle for effective assessment of phoneme
parameters and voice characteristics [5].
It is possible to suppose that alternative models of non-orthogonal frequency
transformations of voice acoustic signal in short time intervals allow to obtain more stable
assessments of the formant speech structure and parameters of voice characteristics.
Experimental model of the program product, created on the basis of investigations, for
International Conference on Software Engineering
realization of these tasks was approved in the expert departments of the Ministry of
Justice and the Ministry of Internal Affairs of Ukraine.
Experimental program system for search of involvants of information messages has a
form of a search engine. Thus, the search results are not identical to identification of
person by voice, because they are just results of ranking by the similarity degree of
separate parameters of voice signals.
Using digital records of information messages the EPS performs an automatic calculation
of parameters of voice characteristics and further ranking of these characteristics in the
voice database.
The EPS uses a ranking method by four different criteria. They include:
- calculation of similarity of two-dimensional probability density functions curves for
the main tone frequency and arrangement in the spectrum of seven formants,
extracted from the speech recorded in the phonogram;
- calculation of similarity of probability density functions curves for each of these
signs separately;
- calculation of similarity degree for the absolute maxima of formats spectra,
extracted from the speech recorded in the phonogram.
Conclusion
As result of fulfilled work the specialized program system for search of involvants of
anonymous messages in large intergovernmental voice databases should be created.
The program system should have several language localizations English, German,
Ukrainian, Russian and other languages. The possibility for quick modification for adding
the new EU localization should be provided. The system should ensure quick access
(with regulation of access rights) to accumulated distributed databases of voice
messages within the frames of unified standard of the EU countries. Structuring of the
system of voice messages within the frames of distributed databases will allow to realize
access to operative solving of problems, including voice messages analysis systems of
other developers.
International Conference on Software Engineering
50
Investigative part of the work should ensure assessment of efficiency and practical utility
of search for involvants of voice messages in the EU countries databases.
Bibliography
1. [Rubalsky et al, 2014] O. Rubalsky, V. Solovyov, A. Shablja, V. Zhuravel. New tools
to identify the person for voice of Data (Russian) Protection and Security of
Information Systems: Proceedings of the 3rd International Scientific Conference (sity.
Lviv, 05 - June 6, 2014). - Lviv: - p.p.. 110 - 112.
2. [Mallat, 1999] Mallat S. A wavelet tour of signal processing, Courant Institute, New
York University, 1999, 671 pp.
3. [Solovyov, 2014] V. Solovyov, Y. Byelozorova: Multifractal approach in pattern
recognition of an announcer’s voice. Polish Academy of Sciences University of
Engineering and Economics in Rzeszów, Teka, Vol. 14, no 2, p.p. 164-170, 2014.
4. [Solovyov, 2013] V. Solovyov. Using multifractals to study sound files (Russian).
Visnik of the Volodymyr Dahl East Ukrainian national university. Vol 9 (151). p. p.
281-287, 2013.
5. [Solovyov, 2013] V. Solovyov. Using the fractal dimension of audio files in the
problem of segmenting the audio file (Russian). Visnik of the Volodymyr Dahl East
Ukrainian national university. Vol 5 (194). p. p. 165-169, 2013.
Authors' Information
Serhii Zybin DSc, Associate Professor, Head of Software Engineering
Department,National Aviation University, Kyiv, Ukraine.
E-mail: zysv@ukr.net
Major Fields of Scientific Research: game theory, cloud computing,
cyberspace, distributed nets, software engineering, information security
Yana Bielozorova Senior Lecturer of Software Engineering
Department, National Aviation University, Kyiv, Ukraine.
E-mail: bryukhanova.ya@gmail.com
Major Fields of Scientific Research: Speech Recognition Models,
Wavelet analysis, Software Architecture
International Conference on Software Engineering
BAYESIAN REASONING AND MACHINE LEARNING
Michael Gr. Voskoglou
Abstract: Although the Bayes’ theorem is a straightforward consequence of the equation
calculating the value of a conditional probability, Bayesian reasoning has been proved to
be very important for the whole science. The present work discusses applications of the
Bayes’ theorem to everyday life situations and its importance for Machine Learning.
Keywords: Conditional Probabilities, Bayes’ Theorem, Artificial Intelligence, Machine
Learning.
ITHEA Keywords: G Mathematics and Computing.
Conference and topic: International Software Engineering Conference, National
Aviation University, Kyiv, Ukraine, 12.04 14.04, 2021.
Introduction
Artificial Intelligence (AI) is the branch of Computer Science focusing on the theory and
practice of creating intelligent machines which work and react like humans, i.e. being
able to think, hear, talk, walk and even feel [Mitchell, 2019]. AI, introduced by John
McCarthy in 1956 [Moor, 2006], is a synthesis of ideas from mathematics, engineering,
technology and science which has the potential to generate enormous benefits to the
human society.
Probability and Statistics are among the main mathematical tools used in AI applications
in which uncertainty plays an important role. Both of them, however, have been
developed on the basis of the principles of the bivalent logic. Consequently, they are able
to tackle effectively only the cases of uncertainty which are due to randomness and note
those due to imprecision. In cases of imprecision the Zadeh’s infinite-valued fuzzy logic
and other related to it theories [Voskoglou, 2019] come to bridge the existing gap.
As we shall see here, however, the Bayes’ theorem calculating the value of conditional
probabilities introduces a kind of multi-valued logic tackling the existing due to
imprecision uncertainty in a way analogous to fuzzy logic. The present work discusses
International Conference on Software Engineering
52
the importance of Bayesian reasoning for Machine Learning, i.e. the branch of AI that
refers to any computer program that can “learn” by itself from a training data set.
The Bayes Theorem
Let A and B be two intersecting events. Then it is straightforward to check [Schuler &
Lipschutz, 2010] that the conditional probability for the event A to happen when the event
B has already happened is calculated by
P(A B)
P(A/B)= P(B)
(1)
In the same way one finds that
P(A B)
P(B/A)= P(A)
, or
P(A B)
= P(B/A) P(A).
Replacing the value of
P(A B)
to (1) one finds that
P(B/A)P(A)
P(A/B)= P(B)
(2).
Equation (2), which calculates the conditional probability P(A/B) with the help of the
inverse in time conditional probability P(B/A), the prior probability P(A) and the posterior
probability P(B), is known as the Bayes’ theorem (or rule, or law). In other words, the
Bayes theorem calculates the probability of an event based on prior knowledge of
conditions related to that event.
The value of the prior probability P(A) is fixed before the experiment, whereas the value
of the posterior probability is calculated with the help of the experiment’s data. Usually,
however, there exists an uncertainty about the exact value of P(A). In such cases,
considering all the possible values of P(A), we obtain through the Bayes’ theorem
different values for the conditional probability P(A/B). Therefore, the Bayes’ theorem
introduces a kind of multi-valued logic tackling the existing, due to the imprecision of the
value of the prior probability, uncertainty. In other words, one could argue that Bayesian
Reasoning constitutes an interface between bivalent and FL.
The Bayes’ theorem was first appeared in the work “An Essay towards a Problem in the
Doctrine of Chances” of the 18th century British mathematician and theologian Thomas
International Conference on Software Engineering
Bayes. This essay was published by Richard Price in 1763, after the Bayes’ death, in the
“Philosophical Transactions of the Royal Society of London”. The famous French
mathematician Laplace (1749-1827), independently from Bayes, pioneered and
popularized the Bayesian probabilities.
When applied in practice, the Bayes’ theorem may have several interpretations. In social
sciences, for example, it describes how a degree of belief expressed as a probability
P(A) is rationally changed according to the availability of related evidence. In that case
the probabilities involved in the Bayes’ theorem are frequently referred as Bayesian
probabilities, although, mathematically speaking, Bayesian and conditional probabilities
are actually the same thing. The outcomes of Bayesian reasoning, however, are not
always compatible to the common beliefs. This will be illustrated here by the following
two timely examples, due to the current COVID-19 pandemic, concerning the creditability
of the viruses’ diagnostic tests.
Example 1: Statistical data show that 2% of the inhabitants of a country have been
infected by a dangerous virus. Mr. X, who has not any symptoms of the corresponding
disease, makes a diagnostic test, the statistical accuracy of which is 97%. The test is
positive. What is the probability for Mr. X to be a carrier of the virus?
Solution: Consider the events A: The subject is a carrier of the virus, and B: The test is
positive. On the basis of the given data it turns out that P(A)=0.02 and P(B/A)=0.97.
Further, among 100 inhabitants of the country, 2 on average are carriers and 98 are non-
carriers of the virus. Assuming that all those people make the test, we should have on
average 2x97%=1.94 positive tests from the carriers and 98x3%=2.94 positive tests from
the non-carriers of the virus, i.e.4.88 in total positive tests. Therefore, P(B)=0.488.
Replacing the values of P(A), P(B/A) and P(B) in equation (2) one finds that
P(A/B)≈0.398. Therefore, the probability for Mr. X to be a carrier of the virus is only
39.8% and not 97%, as it could be thought through a first, rough estimation!
Example 2: Assume now that Mr. X has some suspicious symptoms of the corresponding
disease and that 85% of the people presenting such symptoms have been infected by
International Conference on Software Engineering
54
the virus. Mr. X makes the test, which is positive. What is now the probability for Mr. X to
be a carrier of the virus?
Solution: Let A and B be the events defined in Example 1. Here we have that P(A)=0.85
and P(B/A)=0.97. Further, assuming that 100 people who have suspicious symptoms
make the test, we should have on average 85x97%=82.45 positive tests from the carriers
and 15x0.3% =0.45 from the non-carriers of the virus, i.e. 82.9 in total positive tests.
Therefore, P(B)=0.829. Replacing the values of P(A), P(B/A) and P(B) in equation (2)
one finds that P(A/B)≈0.995. In this case, therefore, the probability for Mr. X to be a
carrier of the virus is 99.5%, i.e. exceeds the statistical accuracy of the test!
In general, the sensitivity of the solution is great, depending on the values of the prior
probability P(A). The greater the value of P(A), the higher the creditability of the test.
The Importance of Bayesian Reasoning for Machine Learning
Although the Bayes’ theorem is a straightforward consequence of equation (1)
calculating the value of a conditional probability, Bayesian reasoning has been proved to
be very important for everyday life situations and for the whole science [Athanassopoulos
& Voskoglou, 2020]. Recent researches give evidence that even most of the
mechanisms under which the human brain works are Bayesian [Bertsch McGrayne,
2012]. Consequently, Bayesian reasoning becomes very useful for AI, which focuses on
the design and construction of machines that mimic the human behavior.
Recently researchers have used ML techniques to develop through the Internet a new
generation of web-based smart learning systems (SLS’s) for various educational tasks. A
SLS is actually a knowledge - based software used for learning and acting as an
intelligent tutor in real teaching and training situations. Such systems have the ability of
reasoning and of providing inferences and recommendations by using heuristic,
interactive and symbolic processing and by producing results from the big data analytics
[Voskoglou & Salem, 2020]. The smart systems of AI are supplied with Bayesian
algorithms in order to be able to recognize the corresponding structures and to make
autonomous decisions.
International Conference on Software Engineering
The physicist and Nobel Prize winner John Mather has already expressed his
uneasiness about the possibility that the Bayesian machines could become too smart in
future, so that to make humans to look useless (http://edge.org/response-detail/26871)!.
Conclusion
The discussion performed in this work leads to the conclusion that Sir Harold Jeffreys
(1891-1989), a British mathematician who introduced the concept of the Bayesian
algorithm and played an important role in the revival of the Bayesian view of probability,
has uprightly characterized the Bayes’ theorem as the “Pythagorean theorem of
Probability theory” (Jeffreys, 1973).
References
1. [Athanassopoulos & Voskoglou, 2020] Athanassopoulos, E. & Voskoglou, M.Gr., A
Philosophical Treatise on the Connection of Scientific Reasoning with Fuzzy Logic,
Mathematics, 8, article 875, 2020.
2. [Bertsch McGrayne 2012] Bertsch McGrayne, S., The Theory that would not die,
Yale University Press, New Haven and London, 2012.
3. [Jeffreys 1973] Jeffreys, H., Scientific Inference, 3d Edition, Cambridge University
Press, UK, 1973.
4. [Mitchell, 2019] Mitchell, M., Artificial Intelligence: A Guide for Thinking Humans,
Parrar, Straus and Gtraux: NY, USA, 2019.
5. [Moor, 2006] Moor, J., The Dartmouth college artificial intelligence conference: The
next fifty years, AI Magazine, 27, 8791, 2006.
6. [Schuler & Lipschutz, 2010] Schuler, J. & Lipschutz, S., Schaum’s Outline of
Probability, 2nd Edition, McGraw-Hill, NY, USA, 2010.
7. [Voskoglou, 2019], Voskoglou, M.Gr., Generalizations of Fuzzy Sets and Relative
Theories, in Voskoglou, M.Gr. (Ed.), An Essential Guide to Fuzzy Systems, pp. 345-
368, Nova Science Publishers, NY, USA, 2019.
8. [Voskoglou & Salem, 2020], Voskoglou, M.Gr & Salem, A.-B.M., Benefits and
Limitations of the Artificial with Respect to the Traditional Learning of Mathematics,
Mathematics, 8, Article 611, 2020
International Conference on Software Engineering
56
Author’s Information
Michael Gr. Voskoglou Graduate Technological Educational Institute
(TEI) of Western Greece; Professor Emeritus of Mathematical Sciences;
26334 Patras, Greece e-mail: mvoskoglou@gmail.com
Major Fields of Scientific Research: Algebra, Artificial Intelligence,
Fuzzy Logic, Markov Chains, Mathematics Education
International Conference on Software Engineering
UKRVECTŌRĒS: AN NLU-POWERED TOOL FOR KNOWLEDGE DISCOVERY,
CLASSIFICATION, DIAGNOSTICS & PREDICTION
Vitalii Velychko, Kyrylo Malakhov, Oleksandr Shchurov
Abstract. The given paper presents a free and open source toolkit which aim is to
quickly deploy web services handling pre-trained distributional semantic models (static
word embeddings). It completes the pipeline between training such models and sharing
the results to the general public. Our web service, UkrVectōrēs, provides all the
necessary routines for online access, sharing and querying to the trained models via
modern Web application, built with Angular, Flask and Docker platforms. We also
describe a demo setup of the UkrVectōrēs toolkit, featuring several efficient models
(“WhiteBook” model, trained using a dataset – the book «The White Book of Physical and
Rehabilitation Medicine» Ukrainian version; “Fiction” model, trained using a dataset –
Ukrainian fiction literature) for Ukrainian.
Key words: Computational Linguistics, Distributional Semantics, Word Embeddings,
Natural Language Processing, Machine Learning, Word2Vec, Visualization, Knowledge
Discovery.
ITHEA Keywords: H.3.3 INFORMATION SEARCH AND RETRIEVAL, H.5.1
MULTIMEDIA INFORMATION SYSTEMS.
Conference and topic: Software Engineering Application Domains: Databases and
Knowledge Bases.
Introduction
In this paper we present UkrVectōrēs [docsim, 2021], a free and open-source toolkit to
deploy web services implementing distributional semantic models, primarily static word
embeddings. UkrVectōrēs is an NLU-Powered (Natural language understanding) tool for
knowledge discovery, classification, diagnostics and prediction.
Distributional semantic models are well established in the field of computational
linguistics and have been here for decades [Turney, 2010]. However, recently they
received substantially growing attention. The main reason for this is a possibility to
employ artificial neural networks trained on small/large/huge corpora to learn low-
International Conference on Software Engineering
58
dimensional distributional vectors for words - word embeddings. The most well-known
tools in this field now are word2vec, fastText built on top of the Skip-Gram and CBOW
algorithms, which allows fast training on huge amounts of raw linguistic data [Mikolov,
2013]. Distributional static word embeddings are now arguably the most popular way to
computationally handle lexical semantics. In such models, every word is represented with
a dense float vector: the number of dimensions is usually in the order of hundreds, and
may vary depending on the purpose of the model [Kutuzov, 2016]. Static word
embeddings represent meaning of words, and can be use in almost any linguistic task:
named entity recognition, sentiment analysis, machine translation, corpora comparison,
compute semantic similarity between entities (such as words, terms and multi-word
expressions) etc. Approaches implemented in word2vec, fastText and other similar tools
are being extensively studied and tested in application to the English language [Baroni,
2016]. However, for many other languages the surface is barely scratched. Thus, it is
important to facilitate research in this field and to provide access to relevant services for
various linguistic communities, in particular Ukrainian linguistic community. UkrVectōrēs
service creation were inspired by WebVectors [WebVectors, 2021], [Kutuzov, 2016] and
RusVectōrēs [Kutuzov, 2016], [RusVectores, 2021] Web services.
UkrVectōrēs deployment basics with Docker Compose
The UkrVectōrēs toolkit is a web interface between distributional semantic models and a
user. Under the hood, we use Gensim open source Python library [Gensim, 2021] which
deals with models’ operations. The user interface is implemented in Typescript (Angular
web framework) and runs on top of the regular Nginx HTTP server or as a standalone
service (using uWSGI application server). It communicates with Gensim (functioning as a
daemon with our wrapper) via REST API (using particular JSON model), sending front-
end queries and receiving back models’ answers. Such architecture allows fast
simultaneous processing of multiple users querying multiple models over network.
Models themselves are permanently stored in memory, eliminating time-consuming stage
of loading them from hard drive every time there is a need to process a query.
International Conference on Software Engineering
UkrVectōrēs can be useful in a very common situation when one has trained a
distributional semantics model for one’s particular corpus or language, but then there is a
need to demonstrate one’s results to general public.
To simplify the application deployment process as much as possible, it was decided to
use the Docker Compose [Docker, 2021] technology (Docker Compose is a tool for
defining and running multi-container Docker applications). As UkrVectōrēs application
grow more complex, we may find significant benefit in running some services in separate
containers. Splitting up application into multiple containers allows us to better isolate and
maintain key services, providing a more modular and secure approach to application
management. Each service can be packaged with the operating environment and tools it
specifically needs to run, and each service can be limited to the minimum system
resources necessary to perform its task. The benefits of multi-container applications
compound as the complexity of the application grows. Because each service can be
updated independently, larger applications can be developed and maintained by
separate teams, each free to work in a way that best supports their service. There are
two containers: Nginx container for accepting external requests; Flask container for
REST API endpoints. The Building and running process under UNIX (Linux or MacOS)
with Docker Compose is as follows:
Clone from UkrVectōrēs GitHub repository:
$ git clone https://github.com/malakhovks/docsim.git
Put your model(s) in ‘server/models’ sub-directory of UkrVectōrēs;
Change configuration files, stating the paths to your models;
Optionally change other settings via ‘config.models.simple.json’ file;
Build and run a Docker image from Dockerfile-nginx, Dockerfile-docsim with
docker-compose:
$ docker-compose up -d
Open Web browser and navigate to UkrVectōrēs Web application.
International Conference on Software Engineering
60
You can preview a completed latest version of the UkrVectōrēs Web service with several
efficient models (“WhiteBook” model, trained using a dataset – the book «The White
Book of Physical and Rehabilitation Medicine» Ukrainian version; “Fiction” model, trained
using a dataset Ukrainian fiction literature) for Ukrainian here: without ssl certificate
http://ukrvectores.ml/; with self-singed ssl certificate https://test.ulif.org.ua:51043.
UkrVectōrēs features
Immediately after that you can interact with the loaded model via UkrVectōrēs Web
application. From a user’s point of view, UkrVectōrēs is a kind of “cognitive-semantic
calculator” which operates on relations between entities (words and terms) in
distributional models. The online toolkit UkrVectōrēs covers the following elements of
distributional analysis:
calculate semantic similarity between pairs of words;
find words semantically closest to the query word;
apply simple algebraic operations to word vectors (addition, subtraction, finding
average vector for a group of words and distances to this average value);
draw semantic maps (using the TensorFlow’ TensorBoard Embedding Projector)
of relations between input words (it is useful to explore clusters and oppositions,
or to test your hypotheses about them);
get the raw vectors (arrays of real values) and their visualizations for words in the
chosen model;
download default models;
use other prognostic models distributive semantics freely distributed, by adjusting
the configuration file.
Another feature of the service is the possibility to use several models simultaneously. If
several models are enumerated in the configuration file, the UkrVectōrēs daemon loads
all of them. UkrVectōrēs graphical user interface is multi-lingual (English and Ukrainian).
Extending the interface with another language is implemented with Angular
internationalization (i18n) feature.
International Conference on Software Engineering
Results and Discussion
Natural Language Processing is the automatic processing of text written in natural
(human) languages (English, French, Ukrainian, etc.), as opposed to artificial languages
such as programming languages, to try to “understand” it. It is also known as
Computational Linguistics or Natural Language Engineering. NLP encompasses a wide
range of tasks, from low-level tasks, such as segmenting text into sentences and words,
to high-level complex applications such as semantic annotation (semantic similarity
between words and multi-word expressions) and data/knowledge mining (automatic
creation of thesauri and bilingual dictionaries, creating semantic maps of different subject
domains). Distributional word embeddings are now arguably the most popular way to
computationally handle lexical semantics. The UkrVectōrēs – a free and open source
toolkit which aim is to quickly deploy web services handling pre-trained distributional
semantic models (static word embeddings). It completes the pipeline between training
such models and sharing the results to the general public. The UkrVectōrēs provides all
the necessary routines for online access, sharing and querying to the trained models via
modern Web application, built with Angular, Flask and Docker platforms.
Conclusion
The main aim of UkrVectōrēs is to quickly deploy web services processing queries to
distributional semantic models, independently of a particular language. It allows to make
complex linguistic resources available to wide audience in almost no time. Additionally,
another major cases for using UkrVectōrēs Web service is to help in the study and
discovery the of arbitrary domains. In particular, using the “WhiteBook” distributional
semantic model (trained using a dataset the book «The White Book of Physical and
Rehabilitation Medicine» Ukrainian version) is intended for training specialists in physical
and rehabilitation medicine. The authors plan to continue adding new features aiming at
better understanding of embedding models, including sentence similarities, documents
similarities and analysis of correlations between dierent kind of models (to changeover
International Conference on Software Engineering
62
from distributed word representations to distributed term representations or term
embeddings).
Bibliography
1. [docsim, 2021] UkrVectōrēs/docsim – an NLU-powered tool for knowledge discovery,
classification, diagnostics and prediction. Entities similarity tool.
https://github.com/malakhovks/docsim.
2. [Turney, 2010] Turney P.D., Pantel P., et al. From frequency to meaning: vector
space models of semantics. J. Artif. Intell. Res., 2010 37(1), pp. 141188.
3. [Mikolov, 2013] Mikolov T., Sutskever I., Chen K., Corrado G.S., Dean J. Distributed
representations of words and phrases and their compositionality. Adv. Neural Inf.
Process. Syst., 2013 26, pp. 31113119.
4. [Kutuzov, 2016] Kutuzov A., Kuzmenko E. WebVectors: A Toolkit for Building Web
Interfaces for Vector Semantic Models. In: Ignatov D. et al. (eds) Analysis of Images,
Social Networks and Texts. AIST 2016. Communications in Computer and
Information Science, vol 661. Springer, Cham.
5. [Baroni, 2016] Baroni M., Dinu G., Kruszewski G. Don’t count, predict! a systematic
comparison of context-counting vs. context-predicting semantic vectors. In:
Proceedings of the 52nd Annual Meeting of the Association for Computational
Linguistics, vol. 1, 2014.
6. [WebVectors, 2021] WebVectors: word embeddings online.
http://vectors.nlpl.eu/explore/embeddings/en/.
7. [RusVectores, 2021] RusVectōrēs: word embeddings for Russian online
https://rusvectores.org/en/.
8. [Gensim, 2021] Gensim Topic modelling for humans.
https://radimrehurek.com/gensim.
9. [Docker, 2021] Overview of Docker Compose tool.
https://docs.docker.com/compose/.
Authors' Information
Vitalii Velychko PhD, Senior researcher. V.M. Glushkov Institute of
Cybernetics of NAS of Ukraine, e-mail: aduisukr@gmail.com
Major Fields of Scientific Research: Artificial Intelligence, NLP
International Conference on Software Engineering
Kyrylo Malakhov Research fellow in IT. V.M. Glushkov Institute of
Cybernetics of NAS of Ukraine, e-mail: malakhovks@nas.gov.ua
Major Fields of Scientific Research: Artificial Intelligence, Computational
linguistics, Distributional semantics
Oleksandr Shchurov Junior researcher. V.M. Glushkov Institute of
Cybernetics of NAS of Ukraine, e-mail: alexlug89@gmail.com
Major Fields of Scientific Research: Artificial Intelligence, NLP
International Conference on Software Engineering
64
MULTILINGUAL TEXT CLASSIFICATION USING TRANSFORMER APPROACH
Larysa Chala, Mykola Dudnik, Mariia Sokolovskaya
Abstract: Recently, experts have been developing methods for processing multilingual
texts using machine learning methods. This article proposes a method that allows one
model to support 100 languages for processing and analyzing text streams. The results
of an experiment on text classification using this model are presented, which confirm its
effectiveness.
Keywords: Deep Learning, Artificial Intelligence, Natural Language Processing.
ITHEA Keywords: D.2.0 Genera
Introduction
As you may know there are two problems in the classification of multilingual texts: the
first one is a linear growth of computational costs with an increasing of number of
languages; the second one is the lack of ability in the model to transfer knowledge
between different languages (as a rule, they are formed only for one language and
cannot be directly applied to another language). At the same time, many different
languages are used on the network (for example, Facebook users publish content in
more than 160 languages).
Existing solutions to these problems are based on building the model, as well as the
entire architecture associated with it, for each language that needs to be worked with.
The approach proposed in this article assumes that one model can be used for 100
languages, which simplifies the server architecture of the word processing system and
provides further support for this solution.
Basic definitions
As you may know there are two problems in the classification of multilingual texts: the
first one is a linear growth of computational costs with an increasing of number of
languages; the second one is the lack of ability in the model to transfer knowledge
between different languages (as a rule, they are formed only for one language and
International Conference on Software Engineering
cannot be directly applied to another language). At the same time, many different
languages are used on the network (for example, Facebook users publish content in
more than 160 languages).
Existing solutions to these problems are based on building the model, as well as the
entire architecture associated with it, for each language that needs to be worked with.
The approach proposed in this article assumes that one model can be used for 100
languages, which simplifies the server architecture of the word processing system and
provides further support for this solution.
Proposed experiment and its results
In the course of research, a news headline classifier was created based on the XLM-R.
This classifier is built using LSTM layers and is a deep neural network. The public kaggle
dataset [4] was used as the training dataset.
The textual dataset on which the XLM-R is trained is not marked up, it is a set of texts
and nothing more. The use of languages with limited resources means that there are
practically no texts for the preliminary training of language models. In our experiment, we
used a tagged dataset consisting of text pairs “news headlines - news categories”.
The results of the experiment showed that when training a classifier in one language
(using XLM-R for vectorization), it will also work well with other languages. In our case,
the model trained in English showed an accuracy of 0.87 on a test set with texts in
French (Fig. 1).
Figure 1. Results of model evaluation
Conclusion
Using the XLM-R model for text vectorization is an important step towards building
multilingual text classifiers.
International Conference on Software Engineering
66
Possible applications using XLM-R technology include serving highly accurate rules-
breaking inappropriate content detection models across a wide range of languages. The
proposed approach makes it possible to implement the principle: "one model for many
languages", in contrast to the existing one, which creates one model for each language
[5].
"One model for many languages" is a new approach that became available after the
emergence of the transformer architecture and the mechanism of self-attention
embedded in them. This approach allows using one model for text vectorization for all
languages that participated in its training.
To sum up, this approach will simplify the further launch of high-performance products in
several languages at the same time. In addition, the approach discussed will improve the
performance of multilingual models and systems that use self-taught methods to better
understand languages with limited resources.
Bibliography
1. [Alammar, 2018] Alammar, Jay. The Illustrated Transformer [Blog post]. Visualizing
machine learning one concept at a time. June 27, 2018.
https://jalammar.github.io/illustrated-transformer/ [Accessed 3 March 2020].
2. [Branden, 2020] Branden, Chan. XLM-RoBERTa: The alternative for non-english
NLP [Blog post]. deepset-ai, Bridging the gap between NLP research & industry.
February 17, 2020. https://medium.com/deepset-ai/xlm-roberta-the-multilingual-
alternative-for-non-english-nlp-cf0b889ccbbf [Accessed 5 March 2020].
3. [Rohan, 2020] Rohan, Jagtab. XLM-RoBERTa: Unsupervised Cross-lingual
Representation Learning at Scale [Blog post]. The Startup, Build Something
Awesome. October 18, 2020. https://medium.com/swlh/xlm-roberta-unsupervised-
cross-lingual-representation-learning-at-scale-f9b144d452cb [Accessed 28
November 2020].
4. [Rishabh, 2019] Rishabh, Misra. News Category Dataset. Kaggle.
https://www.kaggle.com/rmisra/news-category-dataset [Accessed 3 April 2020].
5. [Paul, 2020] Paul, Barba. Challenges in Developing Multilingual Language Models in
Natural Language Processing (NLP) [Blog post]. Towards Data Science, A Medium
publication sharing concepts, ideas and codes. October 23, 2020.
https://towardsdatascience.com/challenges-in-developing-multilingual-language-
models-in-natural-language-processing-nlp-f3b2bed64739 [Accessed 3 April 2020].
International Conference on Software Engineering
6. [Martin and Ondřej, 2018] Martin, P., Ondřej, B., Training Tips for the Transformer
Model. Journal The Prague Bulletin of Mathematical Linguistics 110, April 2018, pp.
43-70. doi: 10.2478/pralin-2018-0002 https://arxiv.org/abs/1804.00247.
Authors' Information
Larysa Chala Kharkiv National University of Radio Electronics,
Ukraine. Department of Artificial Intelligence. Associate Professor,
PhD; e-mail: larysa.chala@nure.ua
Major Fields of Scientific Research: Natural Language
Processing, Software Programming, Machine Learning
Mykola Dudnik Kharkiv National University of Radio Electronics,
Ukraine. Department of Artificial Intelligence. Undergraduate
Student; e-mail: mykola.dudnyk@nure.ua
Major Fields of Scientific Research: Natural Language
Processing, Software Programming, Machine Learning
Mariia Sokolovskaya Kharkiv National University of Radio
Electronics, Ukraine. Department of Artificial Intelligence.
Undergraduate Student; e-mail: mariia.sokolovska@nure.ua
Major Fields of Scientific Research: Natural Language
Processing, Software Programming, Machine Learning
International Conference on Software Engineering
68
EPIDEMIC EFFECTS IN NETWORK INDUSTRIES
Gorbachuk Vasyl, Dunaievskyi Maksym, Syrku Andrii
Abstract: Modern research in the field of global cellular telephony says that technology
reveals the effects of an established base in acceptance (of a new technology) and that
cellular diffusion varies between countries and groups of countries due to technological,
socioeconomic and regulatory factors that influence the diffusion process. The effects of
an established base may come from the extra effects of social dissemination, such as
social learning through uncertainty and regulation pressures. With a severe decrease in
the intensity of use, the epidemic effect is largely absent or at least overshadowed by
other factors, notably consumer heterogeneity. Hence it follows that the driving forces
behind diffusion are constantly falling prices and / or increasing quality of the network
services.
Keywords: network products, network externalities, monopoly, social welfare.
ITHEA Keywords: computer-communication networks, network architecture and design,
network operations, distributed systems, performance of systems.
Introduction
As a rule, existing models cannot empirically differentiate periods of rapid diffusion in
network industry, which are common in most industrialized countries, from critical mass
that does not rely on price reductions or exogenously changing technologies [Gorbachuk,
2013]. Attempts have been made to exclude another potential cause of endogenous
diffusion epidemic effects, in which the rate of penetration increases with constant
intensity of use.
The problem statement
The problem statement is to investigate the possibility of an endogenous and almost
discontinuous diffusion path (new network technology) based on the effect of the
established base (available number of technology users) [Chikrii, Gorbachuk et al.,
2012].
International Conference on Software Engineering
Unanswered question and purpose
An unanswered question is the properties of models which allow to find the critical mass
of technology’s spread and propose appropriate regulation. The purpose of the work is to
strictly justify the regulation in the field of telecommunications [Gorbachuk et al., 2013].
Main results
The main results come from an economy model consisting of two identical groups of
consumers who want to join a certain telecommunications service (for example get a
phone call):
consumers of type
H
highly appreciate joining this service, а
0
consumers of type
L
low [Shy, 2001].
Denote
0p
as fee for connecting to this service. If the actual number of connected (to
the service) consumers is
q
, then the utility function of the consumer of type
L
equals
connectednotconsumer,0
connectedconsumer,pq
L
U
(1)
and the utility function of the consumer of type
H
connectednotconsumer,0
connectedconsumer,pq
H
U
(2)
where
1
measures the importance of service to the consumer of type
H
:
LH UU
.While building demand function for telecommunication services we will use
the assumption about the absence of coordination failure: if every consumer of group
(
H
or
L
) wins from subscription to the service, then all customers of the group will
subscribe to the service.
Critical mass is the minimum number of users required to ensure that at least that
number of consumers will have not negative subscription benefit. In order to organize a
party or weekend trip the organizer must convince potential participants that it will be for
sure attended by a certain minimum number of people, which will in fact mean even more
entrants due to the growing network effects. In telecommunications, critical mass is
International Conference on Software Engineering
70
always a function of market price: an increase in price will mean an increase in critical
mass, and a decrease in market price will reduce critical mass, since under a lower price
consumers will be satisfied with a smaller network size. With this concept in mind, a
former student of one of the authors developed the Djuice project, which earned Kyivstar
more than two billion U.S. dollars over five years.
By the 1980s, most countries had a monopoly market structure in the field of
telecommunications. For example, in Ukraine such a monopoly firm was called
Ukrtelecom, in Estonia Estonia Telecom. In Israel, until the 1960s, a similar firm also
provided postal services.
During the 1980s, governments began to realize that monopoly telecommunications
market structures, which were considered natural monopolies, distorted industry markets.
The major event leading to competition in this area was the division of AT&T USA into 7
regional telephone companies in 1982, as well as the creation of MCI and SPRINT as
major competitors in the intercity and international markets. In the 1980s, regulators were
discussing three basic issues: knowing that many users (of type
H
) are already
connected to an existing monopoly telecommunications provider, can social well-being
be improved by allowing the new operator to connect other consumers (of type
L
) to the
network; whether the new operator entering the market will have profit; when the entry of
new providers is socially desirable, how to avoid dragging of the existing monopoly into
predatory pricing to attract more consumers and thus narrow the potential market for a
new firm [Gorbachuk, 2014].
Proposition 1. Entry into the telecommunications industry increases the utility of already
connected customers and leaves unchanged the utility of newly connected consumers,
increases the profits of a new firm and leaves the profits of an existing firm unchanged.
When, in addition to the connection service market, the market for the flow of services
(telephone calls) after the connection of consumers of type
L
at the entry of a new firm
is also taken into account, the existing firm will suffer a decrease in profits, but social
welfare will increase as a result of the price decrease. The analysis of the
International Conference on Software Engineering
telecommunications industry is based on the fact that the utility of the consumer from the
communication service is enhanced when others join the service. Let us denote
]1,0[x
the reluctance of the consumer to pay for the service: the greater the value of
x
, the less willing to pay the consumer is. The utility of consumer of type
x
will define as
connectednot consumer
connectedconsumer
,0
,p
e
q)x1(
x
U
(3)
where
e
q
expected number of consumers who join this network. As this consumer
utility increases with increasing
e
q
, have network externalities.
Proposition 2. A monopoly telephone company maximizes its profits by setting its cost
per connection so that the number of users exceeds half of all existing customers and
that unconnected consumers exist.
Proposition 3. With increasing of the total consumers number
the monopoly price and
the utility of the connected users increase proportionally, monopoly profit increases
quadratically.
Conclusion
The main conclusion is that the entry of a new firm into the telecommunications industry
increases the utility of already connected consumers and leaves unchanged the utility of
newly connected consumers. A monopoly network company maximizes its profits by
setting its price per connection so that the number of users exceeds half of all customers
available, but not all customers subscribe to the network service. Questions that need
further investigation are: whether exogenous changes are responsible for total diffusion
(without critical mass); whether there is an element of endogenous diffusion.
Bibliography
1. [Shy, 2002] Shy O. The economics of network industries. Cambridge: Cambridge
University Press, 2001. 315 p.
2. [Chikrii, Gorbachuk et al., 2012] Chikrii A., Denisova N., Gorbachuk V., Gromaszek
K., Krivonos Y., Lytvynenko V., Matychyn I., Osypenko V., Smailova S., Wojcik W.
International Conference on Software Engineering
72
Current problems in information and computational technologies. V. 2. W. Wojcik, J.
Sikora (eds.) Lublin, Poland: Politechnika Lubelska, 2012. 196 p.
3. [Gorbachuk, 2013] Gorbachuk V. The ways of coping with market imperfections in
the telecommunications industry. Odesa National University Herald. Economy. 2013.
V. 18. 3 (1). P. 84-87. (In Ukrainian).
4. [Gorbachuk et al., 2013] Gorbachuk V., Chumakov B. A computable critical mass on
market of new network product. Transport systems and logistics. Chisinau, Moldova:
Academia de Transporturi, Informatică şi Comunicaţii, 2013. P. 272281.
5. [Gorbachuk, 2014] Gorbachuk V.M. Equilibria of international public goods.
Education and science and their role in social and industrial progress of society.
Kyiv: Alexander von Humboldt Foundation, 2014. P. 1314.
Authors' Information
Gorbachuk Vasyl V.M.Glushkov Institute of Cybernetics of the National Academy of
Sciences of Ukraine, DSc (Physics and Mathematics), Head of Department, e-mail:
GorbachukVasyl@netscape.net
Major Fields of Scientific Research: Equilibria of Decentralized Systems
Dunaievskyi Maksym V.M.Glushkov Institute of Cybernetics of the National
Academy of Sciences of Ukraine, MSc (Finance), PhD Student, e-mail:
MaxDunaievskyi@gmail.com
Major Fields of Scientific Research: Data Science Techniques
Syrku Andrii V.M.Glushkov Institute of Cybernetics of the National Academy of
Sciences of Ukraine, MSc (Engineering), PhD Student, e-mail: saan@ukr.net
Major Fields of Scientific Research: Smart Sensors, Signal Processing
International Conference on Software Engineering
Тези доповідей
Міжнародної науково-практичної конференції
Інженерія програмного забезпечення
Технічнi редактори Markov K., Чебанюк О.В.
Матеріали друкуються за редакцією авторів.
Підписано до друку.
© Національний авіаційний університет
... As a rule, the ratio p q / is nearing 0 for profitable products; therefore, the sales volume S reaches its maximum, when the aggregate sales volume is nearing m / 2 and exceeding critical mass [26][27][28][29] as follows: ...
Article
Full-text available
The paper examines prerequisites and assumptions of the classical Bass innovation diffusion model with the aim of applying it in modeling of relevant stochastic processes related to the pandemic. The Bass model has proven its versatility and applicability to various environments. A thorough mathematical substantiation of the model properties is presented based on theories of evolutionary equations and stochastic processes for its further development, as well as search for uncertainty parameters and observable variables. The paper provides realistic estimation results of the Bass model parameters for vaccination in Ukraine and Belarus on weekly data of the first half of 2021. Similar studies are suggested for other countries, as well as regions and districts of Ukraine.
Book
Full-text available
The scientific monograph "Digital Economy and Information Society. Selected Issues" is devoted to discussing important issues of science and education in the development of civil society and is relevant today. The monograph considers the following aspects: the forming of the information society; educational and psychological issues of information society development; modern problems of digital economy development. Scholars have represented the results of generalization of theoretical concepts, empirical research and practical experience of modern education and science in the development of digital economy and information society. The authors' scientific research has once again shown that current trends in education and science in the development of the information society reflect the objective need to consider them as an integral part of the digital economy and social system, which plays a special role in human development. Thus, the digital economy is the driving force of the information society development, and requires the participation of all social system’s elements in its forming by means of using modern information and communication and digital technologies and aspects, which have been described by the authors in the monograph and those that continue to be developed and tested. The monograph meets the principles of academic integrity; it is performed at a high scientific level, and can be useful for scientists, economists, educators, students.
Article
The article examines in detail the prerequisites and assumptions of the classical Bass model of innovation diffusion in order to apply it to the modeling of topical stochastic processes related with the pandemic. The Bass model has proven its versatility and applicability to various environments. The article presents a thorough mathematical substantiation of the model properties for the purpose of its further development, search of uncertainty parameters and observed variables, based on the theories of evolutionary equations and stochastic processes. The paper provides realistic results of the Bass model parameter estimation for vaccinations in Belarus and Ukraine on the weekly data of the first half of 2021 and suggests similar studies for other countries as well as regions and districts of Ukraine.
Article
Full-text available
Introduction. Optimization can be applied in developing profitability management tools for a cloud service broker working according to a certain business model. On behalf of the managing telecommunications holding company (telecommunications operator), this broker integrates, aggregates and configures software and data storage services of third-party Internet software vendors. Such a broker receives only fixed commissions from this company, based on the subscription fee, but does not pay royalties to an Internet software vendor and does not receive payments from the sale of service packages. The purpose. The cloud broker faces the problem of limited human resources required to carry out the relevant legal, technical and economic activities. In addition, the broker faces the problem of uncertainty in sales, service prices, the share of resource use, or the risk of losing operational and financial goals. Results. To run a brokerʼs business efficiently, one needs to find services and their bundles that increase profitability and reduce financial risk by solving certain optimization problems. Information on such services is needed to support negotiations on fixed and variable commissions, as well as to prioritize services and their packages to be provided. Thus, for the cloud services broker, both profitability management tools and services portfolio development tools are useful. In general, a cloud service broker is an organization that negotiates the relationships between cloud service clients and Internet software vendors. Cloud broker can be created on the basis of different business models regarding the type of service (platform, infrastructure, software), type of clients (enterprise, household), functions performed (identity management, accounting, billing, location, etc.), the degree of rebranding, measures of aggregation of services and other criteria. Conclusions. Different cloud brokers have different attitudes to choice of important solutions for their businesses. Solutions can relate to pricing, capacity planning and utilization in combination with service quality, security, scalability and other issues. Keywords: optimization, portfolio, uncertainty, Boolean variables, revenue generation.
Data
Full-text available
People rely on daily news to know what is happening around the world. In today’s world, when the proliferation of fake news is rampant, having a large-scale and high-quality source of authentic news articles with the published category information would be valuable to learning authentic news’ Natural Language syntax and semantics. We present News Category Dataset that contains around 200K news headlines from the year 2012 to 2018 obtained from HuffPost. To make it more useful, we have included the source links of the news articles so that more data can be extracted as needed. In this paper, we describe various details about the dataset and potential use cases where it can be used.
Conference Paper
Full-text available
Several new and revised ISO standards will be published in 2016/17 that define the basic terms and concepts of usability (ISO 9241-11), give guidance on processes and outcomes of human-centred design (ISO 9241-220), provide examples of measures that can be used in usability evaluation (ISO/IEC 25022 and 25023) and define what should be included in usability evaluation reports for usability tests, inspections and surveys (ISO/IEC 25066). The paper explains some of the new content and how it can be used.
Article
Full-text available
For a purpose of suitable system product development and acquisition successfully, it is extremely important to specify the quality requirement for target system during system design phase, and developing or selecting a most suitable product finally from among the alternative candidate products. However, the quality requirements analysis is a very difficult matter because it is non-formulaic and demands much of product customer’s technical perceptiveness. Traditionally, analysis of quality requirements of a product has been conducted through questionnaires or interviews with customers based on survey investigator’s personal experience. This kind of lack of structure and incompleteness in the traditional approach introduced missing requirements and errors in prioritizing requirements to implement in the requirement analysis of a product. On the other hand, we have been working on developing the techniques (SQuaRE) for quality requirements and evaluation for system and software product for a long time in ISO/IECJTC1/SC7WG 6. For the purpose of specify the quality requirement of system product, the method of quantitative quality requirement definitions based on ISO/IEC9126 [1] quality model that includes six characteristics is widely recognized. However, independency among each quality characteristic is not sure and the suitability of method by using these six quality characteristics for quality requirement is not certified statistically. In this paper, we propose the concept of quality requirements definition method based on the SQuaRE quality models define in ISO/IEC25030 [2,3]. Also, this paper proposes the result of verification about effectiveness of quantitative quality requirement definition approach from the view point of six quality characteristics described in ISO/IEC9126.