<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Archiving and Interchange DTD v1.0 20120330//EN" "JATS-archivearticle1.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink">
  <front>
    <journal-meta>
      <journal-title-group>
        <journal-title>ORCID:</journal-title>
      </journal-title-group>
    </journal-meta>
    <article-meta>
      <title-group>
        <article-title>Methodology of Formation of the Individual Study Plan of the Student Based on the Graph Model of the Dependence of Disciplines</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <string-name>Svitlana Krepych</string-name>
          <xref ref-type="aff" rid="aff1">1</xref>
          <xref ref-type="aff" rid="aff2">2</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Iryna Spivak</string-name>
          <xref ref-type="aff" rid="aff1">1</xref>
          <xref ref-type="aff" rid="aff2">2</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Serhii Spivak</string-name>
          <xref ref-type="aff" rid="aff0">0</xref>
          <xref ref-type="aff" rid="aff2">2</xref>
        </contrib>
        <aff id="aff0">
          <label>0</label>
          <institution>Ternopil Ivan Puluj National Technical University</institution>
          ,
          <addr-line>Ruska str. 56, Ternopil, 46001</addr-line>
          ,
          <country country="UA">Ukraine</country>
        </aff>
        <aff id="aff1">
          <label>1</label>
          <institution>West Ukrainian National University</institution>
          ,
          <addr-line>Lvivska str. 11, Ternopil, 46009</addr-line>
          ,
          <country country="UA">Ukraine</country>
        </aff>
        <aff id="aff2">
          <label>2</label>
          <institution>method, Individual Study Plan, High-Quality Specialists</institution>
          ,
          <addr-line>Personal Skills</addr-line>
        </aff>
      </contrib-group>
      <volume>000</volume>
      <fpage>0</fpage>
      <lpage>0001</lpage>
      <abstract>
        <p>The article considers the problem of forming a personalized study plan for a student in accordance with his academic performance. It is noted that currently, students choose elective subjects at random, based on the principle of mass or a certain "curiosity". It is proposed to develop a method of forming an individual study plan for each individual student, taking as a basis performance indicators for previously studied disciplines. Thus, using the indicated method to form a student's individual educational plan, a higher educational institution will be able to form high-quality specialists, relying on their personal skills and educational achievements. Educational Achievements MoMLeT+DS 2023: 5th International Workshop on Modern Machine Learning Technologies and Data Science, June 3, 2023, Lviv, Ukraine</p>
      </abstract>
      <kwd-group>
        <kwd>Dependence</kwd>
        <kwd>Keywords1</kwd>
      </kwd-group>
    </article-meta>
  </front>
  <body>
    <sec id="sec-1">
      <title>-</title>
      <p>Indicated</p>
    </sec>
    <sec id="sec-2">
      <title>1. Introduction</title>
      <p>Higher education in various countries has developed rapidly over time. In order to improve the
education of their students, universities changed the teaching methodology, selected a convenient
number of subjects and derived the optimal study time for each discipline.</p>
      <p>
        Considering the events of recent years in the world, in particular with the spread of the COVID-19
pandemic [
        <xref ref-type="bibr" rid="ref1">1</xref>
        ], there is an urgent need to improve distance education. Considering the realities of today's
Ukrainian student, education was also forced to stand on the rails of "war" and adapt to new challenges.
In order to improve the education of their students, universities began to change teaching methods,
select a convenient number of subjects and adapt the teaching of material for maximum effect in
minimum time within each discipline. In most modern universities, there are different types of teaching
methods, from remote to selective form, thereby giving students the opportunity to develop in the
direction in which they see themselves best. Thus, they do not limit themselves in terms of time, as they
can choose a convenient time for the couple or in choosing the academic disciplines they wish to study.
      </p>
      <p>The main idea behind the development of the software system is to create a personalized study plan
for the student according to his performance. In this way, the university will train first-class and highly
specialized professionals, which are so needed in our time. All this will take place with the help of an
algorithm, which will monitor the success of each student and, in accordance with the success of
previously studied disciplines, will build the most optimal educational plan for the next academic
period.</p>
      <p>2023 Copyright for this paper by its authors.</p>
    </sec>
    <sec id="sec-3">
      <title>2. Related works</title>
      <p>Higher education in different countries and even universities has developed rapidly over time. In
order to improve the education of their students, universities changed the teaching methodology,
selected a convenient number of subjects and derived the optimal study time for each discipline.
Speaking about the latter, we want to understand that the student who is studying will get the right
amount of knowledge that will be needed in his future profession. There are different types of methods
of choosing elective subjects and now we will consider it.</p>
      <p>
        Consider the education system of the United States of America. Higher education institutions in this
country have a very flexible curriculum. A student can independently choose the time when he wants
to study, in particular evening or morning hours, choose the form of education - face-to-face, part-time
or distance learning - and there is even such a term as "university without walls". However, the most
important thing in a student's choice about his studies is the choice of disciplines at his own will [
        <xref ref-type="bibr" rid="ref2">2</xref>
        ]. To
choose electives, you must earn a certain number of credits to be able to choose a subject from the list.
However, subordination in the proportionality of professional disciplines and disciplines of free choice
must be observed in order to develop the professional quality of the specialist.
      </p>
      <p>If we consider the system of higher education in Germany, we will notice certain differences. Due
to the recent events in the world, the COVID-19 virus, there was a need for a distance learning system
- this is where the main feature emerged.</p>
      <p>In German educational institutions, there is a new education system that allows you to study at home.
However, this is not the last mission of this product. Thanks to the Stud.ip software system, students
have the opportunity through their personal account to view pairs online and lecture recordings, the
opportunity to choose their own study schedule and contact teachers directly.</p>
      <p>
        The method of choosing disciplines in this country is implemented through a personalized office of
the university, where a student will be able to choose not only a set of disciplines he wants to study, but
also teachers, lecturers, personally draw up a schedule of classes and communicate with potential
employers [
        <xref ref-type="bibr" rid="ref3">3</xref>
        ].
      </p>
      <p>Speaking for Ukraine, the scheme here is quite simple. Students choose from a selected list of
disciplines that they like. As in any university, there is a list of professional disciplines that must be
read in the course, as well as optional professional and free cycles.</p>
      <p>
        According to the Law of Ukraine "On Higher Education" (Article 62), persons studying in
institutions of higher education have the right to choose academic disciplines within the limits provided
for by the relevant educational program and curriculum, in the amount of at least 25% of the total
number of ECTS credits determined for a certain level of higher education [
        <xref ref-type="bibr" rid="ref4">4</xref>
        ].
      </p>
      <p>Ukrainian higher education institutions use the ECTS system. ECTS, or the European Credit
Transfer and Accumulation System, is a system that was created to ensure a unified interstate training
evaluation procedure, a system for measuring and comparing learning outcomes, their academic
recognition and transfer from one educational institution to another.</p>
      <p>Let's consider the procedure for choosing disciplines at WUNU (Western Ukrainian National
University). On the basis of curricula, the departments form lists of optional academic disciplines that
can be submitted for the next academic year. Students, in turn, receive a list and make a choice in the
form of writing an application at the department of the university to which they are assigned, or having
access to the electronic system "eUniversity" by logging in there as a student, and choosing the list of
proposed disciplines. Based on the majority of submitted applications or votes left in the "eUniversity"
system, a certain number of optional disciplines is selected for the next academic year.</p>
      <p>Summarizing the above, we can say the following - "The error of any choice is the choice itself."
That is, sometimes we ourselves do not realize, or we do not have enough data and facts to make a
successful choice for ourselves.</p>
    </sec>
    <sec id="sec-4">
      <title>3. Overview of the Research</title>
      <p>Due to the fact that the majority of students randomly choose elective disciplines of a professional
type, it is proposed to automate this system, namely:
- professional disciplines remain mandatory;
- students will be able to choose elective subjects of the free type at their own will;
- selective disciplines of the professional type will be formed on the basis of the method of forming
the individual study plan of the student.</p>
      <p>
        The specified method will work on the basis of a constructed graph with the ratio of compulsory
subjects of the professional cycle and optional subjects. That is, with the help of the specified method,
we will be able to make a personalized schedule of subjects for each student [
        <xref ref-type="bibr" rid="ref5">5</xref>
        ]. Thus, the higher
educational institution will be able to form high-quality specialists, relying on their personal skills and
educational achievements. This method not only optimizes the choice of disciplines and the selection
of the curriculum, but also facilitates the study of higher education students.
      </p>
    </sec>
    <sec id="sec-5">
      <title>4. Proposed method</title>
      <p>
        As a graph model [
        <xref ref-type="bibr" rid="ref6 ref7 ref8">6-8</xref>
        ] of the dependence of the disciplines of the professional cycle of the student's
curriculum, consider an oriented graph G(T, A) , where T  ti , i  0..N – set of graph tops,
А  а ij , i  0..N, j  0..J – set of graph arcs. The dependence graph of the disciplines of the
professional cycle of the student's curriculum G(T, A) is built on the basis of both normative and
selective disciplines. The set of tops T of the graph are all normative and selective disciplines of the
student's professional training cycle. The set of arcs A reflects the connection between the disciplines,
i.e., if there is a transition from the i -th top (discipline) to the j -th top (discipline), then the graph
G(T, A) contains an arc аij that connects these two tops.
      </p>
      <p>
        When modeling the computing process on a graph model, it is assumed that each element of the
model - a top and an arc - is assigned a certain weight [
        <xref ref-type="bibr" rid="ref10 ref11 ref9">9-11</xref>
        ]. Let's assume that each top ti is
characterized by an elementary indicator mi that is directly related to the student's success, that is, the
credit (examination) score given for passing the corresponding і - th discipline. Some tops of the graph
will have the indicated indicator mi  0 , because not all disciplines of the elective cycle will be
assigned to the student's individual study plan. The entered indicators form a set M  mi  on the graph.
      </p>
      <p>In turn, each arc aij which connects the tops ti and t j is characterized by a certain weight, in
particular, the weight coefficient k(i, j) . The specified weight factor is set for each arc in the range
from 0 to 1. Given that each top ti of the graph is a directly specific discipline, normative or selective,
the value of the weight factor k(i, j) will be determined by the conditional influence of the results of
mastering the discipline of the top ti , which is measured by the indicator mi on the discipline of the
selective cycle that contained in the top t j .</p>
      <p>All weighting factors form a set K  k (i, j) on the graph.</p>
      <p>The thus built graph model of the dependence of the disciplines of the professional cycle of the
student's curriculum is an acyclic oriented loaded graph G(T , A, M , K) .</p>
      <p>
        The power of the set М ( М ) is determined by the total number of subjects of the professional
cycle of education - normative and selective [
        <xref ref-type="bibr" rid="ref12">12</xref>
        ].
      </p>
      <p>For the convenience of entering, editing and changing the weights of the arcs of the graph, which
may be caused by the clarification of educational plans or the introduction of new disciplines, it is
advisable to present this indicator in matrix form. So, the weighting coefficients k(i, j) of the graph
model G(T , A, M , K) are represented by a matrix KCon of dimension L  N , where N is the number of
all subjects of the professional cycle of education, and L is the number of only selective subjects of the
second, third and fourth courses of study.</p>
      <p>The matrix KCon (G)  (kij )i1..N, j1..L is a loaded adjacency matrix in which:</p>
      <p>k (i, j), if aij  A,
kij  
0, if aij  A.
(1)
For a loaded adjacency matrix KCon (G)  (kij )i1..N, j1..L , the following condition must be fulfilled:
N N N
 ki1   ki2  ...   kil 1, j  1..L, L  N.</p>
      <p>i1 i1 i1
Consider a matrix GKM whose elements are products kij  mi , i  1...N, j  1...L.
and fourth. However, if for the second year we can immediately choose from the set Srait of maximum
ratings of specific disciplines, then for the third it will be necessary to wait one year of study and
accordingly for the fourth - another year of study. Since the student's individual study plan is built step
by step after the end of each academic year.</p>
      <p>Therefore, there are two ways to solve this situation:
1. After each academic year, update the matrix GKM and the set Srait and act only in the ranges of
values allocated for a certain academic year.</p>
      <p>
        2. After forming the loaded graph G(T, A, M , K) of the dependence of the disciplines of the
professional cycle of the student's curriculum, immediately decompose it, i.e. construct three subgraphs
[
        <xref ref-type="bibr" rid="ref13">13</xref>
        ] G1  G , G2  G and G3  G according to the order of the student's academic year. For each
formed subgraph, we form the corresponding adjacency matrix GKM1  GKM , GKM2  GKM and
GKM3  GKM and the significance set Srait1  Srait , Srait2  Srait and Srait3  Srait .
      </p>
      <p>The next step in the implementation of the method is the selection Х subjects from the significance
set with the maximum value. The number of disciplines for each year of study can be the same or set
manually by the coordinator. The specified number can be specified immediately or specified directly
in the year of formation of a new branch of the student's individual study plan.</p>
      <p>That is, to determine the set of elective subjects for the second year of study, we can match each
student with a set ordered in descending order:</p>
      <p>Srait1  SORT Srait1 
where the operator SORT () arranges the elements of the set in descending order.</p>
      <p>The first X disciplines from the ordered set Srait1 will be included in the student's individual study
plan, the rest are simply zeroed out.</p>
      <p>Now let's move on to modeling the method of building an individual student plan. As you know, the
university has professional disciplines that are necessarily included in the curriculum, and optional ones
that the student chooses on his own.</p>
      <p>In order to combine the student's previous success and the choice of further disciplines, it is
suggested to develop a system of interrelationship of normative and selective disciplines due to the
percentage ratio. These percentages will be calculated based on the student's overall performance for
the course completed.</p>
      <p>For visual perception, with the help of Figure 1, a graphic and with the help of Figure 2, we provide
a tabular representation of the specified interrelationship of disciplines for first-year students when
choosing second-year disciplines.
(2)
(3)
(5)</p>
      <p>In order to form the curriculum for the second year, we will add optional subjects to professional
subjects. Now, the choice of elective course subjects will be influenced by the elective subjects of the
previous academic year. That is, now we get the following table of coefficients (see Fig. 3).</p>
      <p>We complete the modeling process by forming selective disciplines for the fourth year of study. The
algorithm is identical to the previous step: optional subjects of the previous academic year are added to
the normative subjects of the current period and affect the choice of subjects for the fourth year.</p>
    </sec>
    <sec id="sec-6">
      <title>5. Results &amp; Discussion</title>
      <p>To present the results of the system, a conditional group of five students and their simulated
academic performance was created. The following figures will demonstrate the dependence of one
student's success on a professional course to form a set of elective subjects for subsequent courses.</p>
      <p>Figure 5 shows the performance of conditional student Avramenko O. in the first year, and Figure 6
shows Dyriv H.</p>
      <p>Therefore, two optional disciplines from the first year, namely "Web programming" and "Basis of
smart technologies" (Avramenko), "Basis of interaction design" and "Engineering graphics" (Dyriv) are
included in the student's curriculum for the next year and, accordingly , according to the results of the
assessment and examination session, will be evaluated for a certain point. All other optional disciplines
from the first year are "zeroed". From here, based on the performance evaluations of a conditional
student in the second year, we will receive two elective subjects for the third year: "Modern software
development platforms" and "Basis of cloud technologies" (Avramenko) and "Basis of cloud
technologies" and "Data security" (Dyriv) (see Fig. 7-8).</p>
      <p>By analogy with the formation of elective disciplines for the third year, we are forming elective
disciplines for students for the fourth year of study. From here, based on the performance evaluations
of a conditional student in the third year, we will receive two optional subjects for the fourth year:
"Computer graphics" and "Basis of cloud technologies" (Avramenko) and "Data compression methods"
and "Data transmission systems" (Dyriv) (see Fig. 9-10).</p>
      <p>
        The software implementation of the method is given below [
        <xref ref-type="bibr" rid="ref13 ref14">13, 14</xref>
        ]. Figures 11 and 12 show the
progress tables of first-year and third-year students, respectively.
      </p>
      <p>After the grades for all subjects have been posted, the page for viewing the student's calculated
individual plan for the next year will be available.</p>
      <p>Figure 13 shows the page of elective subjects assigned to the respective students, which will be
added to the normative ones in the students' curriculum for the second year.</p>
    </sec>
    <sec id="sec-7">
      <title>6. Conclusions</title>
      <p>The article considers the problem of forming a student's individual study plan, in particular, on the
basis of the elective subjects chosen by the student. It is noted that currently students choose elective
subjects arbitrarily, based on the principle of mass or certain "interest". Quite often, as a result, the
student simply does not like the discipline he has chosen. The authors proposed to make the method to
the selection of elective subjects automated, namely by developing a method of forming an individual
study plan for each individual student, taking as a basis success indicators from previously studied
subjects. From here, the normative part of the subjects will build the basis of the specialist as such, and
a number of elective disciplines formed on the basis of the student's performance indicators will only
improve and increase knowledge in those areas where the student is the best. Thus, using the specified
method of forming a student's individual educational plan, a higher educational institution will be able
to form highly qualified specialists, relying on their personal abilities and educational achievements.</p>
    </sec>
    <sec id="sec-8">
      <title>7. References</title>
    </sec>
  </body>
  <back>
    <ref-list>
      <ref id="ref1">
        <mixed-citation>
          [1]
          <string-name>
            <surname>Vdovichenko</surname>
            <given-names>V.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Vorotniuk</surname>
            <given-names>M.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Gerasemchuk</surname>
            <given-names>S.</given-names>
          </string-name>
          , and others, “
          <article-title>The EU's COVID-</article-title>
          19 Pandemic: Selfishness, Solidarity or Spreading Influence” Analytical Note, Foreign Policy Council,
          <year>2020</year>
          . pp.
          <fpage>4</fpage>
          -
          <lpage>5</lpage>
          .
        </mixed-citation>
      </ref>
      <ref id="ref2">
        <mixed-citation>
          [2]
          <string-name>
            <surname>Stoyka</surname>
            <given-names>O.</given-names>
          </string-name>
          , “
          <article-title>Higher education in the USA and Ukraine: distinctive and common features”</article-title>
          , Scientific Bulletin of the Uzhhorod University,
          <year>2017</year>
          . p.
          <fpage>280</fpage>
        </mixed-citation>
      </ref>
      <ref id="ref3">
        <mixed-citation>
          [3]
          <string-name>
            <surname>Otroshchenko</surname>
            <given-names>L.</given-names>
          </string-name>
          “
          <article-title>Management experience of the system of higher education in Germany”</article-title>
          , autoref.: Kyiv,
          <year>2018</year>
          . 127 p.
        </mixed-citation>
      </ref>
      <ref id="ref4">
        <mixed-citation>
          <article-title>[4] "On higher education: Law of Ukraine dated July 1, 2014 No. 1556-VII", (https://zakon</article-title>
          .rada.gov.ua/laws/show/1556-18)
        </mixed-citation>
      </ref>
      <ref id="ref5">
        <mixed-citation>
          [5]
          <string-name>
            <surname>Spivak</surname>
            <given-names>I.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Krepych</surname>
            <given-names>S.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Spivak</surname>
            <given-names>S.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Fedorov</surname>
            <given-names>O.</given-names>
          </string-name>
          ,
          <article-title>"Approach to estimate the level of influence of motivation on the effectiveness of employees9 depending on their needs"</article-title>
          ,
          <source>3rd International Conference on Advanced Information and Communications Technologies, AICT 2019 - Proceedings 8847890</source>
          , pp.
          <fpage>46</fpage>
          -
          <lpage>49</lpage>
        </mixed-citation>
      </ref>
      <ref id="ref6">
        <mixed-citation>
          [6]
          <string-name>
            <given-names>E.</given-names>
            <surname>Choi</surname>
          </string-name>
          ,
          <string-name>
            <given-names>S.</given-names>
            <surname>Smith</surname>
          </string-name>
          and
          <string-name>
            <given-names>E.</given-names>
            <surname>Young</surname>
          </string-name>
          ,
          <article-title>"Comparing Shallow and Deep Graph Models for Brain Network Analysis"</article-title>
          ,
          <source>IEEE International Conference on Big Data (Big Data)</source>
          ,
          <year>2022</year>
          , pp.
          <fpage>4962</fpage>
          -
          <lpage>4967</lpage>
        </mixed-citation>
      </ref>
      <ref id="ref7">
        <mixed-citation>
          [7]
          <string-name>
            <surname>Harith</surname>
            <given-names>A.</given-names>
          </string-name>
          <string-name>
            <surname>Dawood</surname>
          </string-name>
          ,
          <article-title>"Graph Theory and Cyber Security"</article-title>
          ,
          <source>3rd International Conference on Advanced Computer Science Applications and Technologies</source>
          ,
          <year>2014</year>
          , pp.
          <fpage>90</fpage>
          -
          <lpage>96</lpage>
        </mixed-citation>
      </ref>
      <ref id="ref8">
        <mixed-citation>
          [8]
          <string-name>
            <given-names>P.</given-names>
            <surname>Yanardag</surname>
          </string-name>
          and
          <string-name>
            <given-names>S.</given-names>
            <surname>Vishwanathan</surname>
          </string-name>
          ,
          <article-title>"Deep graph kernels"</article-title>
          ,
          <source>Proceedings of the 21th ACM SIGKDD international conference on knowledge discovery and data mining</source>
          ,
          <year>2015</year>
          , pp.
          <fpage>1365</fpage>
          -
          <lpage>1374</lpage>
          .
        </mixed-citation>
      </ref>
      <ref id="ref9">
        <mixed-citation>
          [9]
          <string-name>
            <given-names>B.</given-names>
            <surname>Scalzo Dees</surname>
          </string-name>
          ,
          <string-name>
            <given-names>Y.</given-names>
            <surname>Xu</surname>
          </string-name>
          ,
          <string-name>
            <given-names>A.</given-names>
            <surname>Constantinides</surname>
          </string-name>
          and
          <string-name>
            <given-names>D.</given-names>
            <surname>Mandic</surname>
          </string-name>
          ,
          <article-title>"Graph Theory for Metro Traffic Modelling"</article-title>
          ,
          <source>International Joint Conference on Neural Networks (IJCNN)</source>
          ,
          <year>2021</year>
          , pp.
          <fpage>1</fpage>
          -
          <lpage>5</lpage>
        </mixed-citation>
      </ref>
      <ref id="ref10">
        <mixed-citation>
          [10]
          <string-name>
            <given-names>S.</given-names>
            <surname>Guze</surname>
          </string-name>
          ,
          <article-title>"Graph theory approach to transportation systems design and optimization"</article-title>
          ,
          <source>TransNav: International Journal on Marine Navigation and Safety of Sea Transportation</source>
          , vol.
          <volume>8</volume>
          ,
          <year>2014</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref11">
        <mixed-citation>
          [11]
          <string-name>
            <given-names>L.</given-names>
            <surname>Stankovic</surname>
          </string-name>
          ,
          <string-name>
            <given-names>D.</given-names>
            <surname>Mandic</surname>
          </string-name>
          ,
          <string-name>
            <given-names>M.</given-names>
            <surname>Dakovic</surname>
          </string-name>
          ,
          <string-name>
            <given-names>M.</given-names>
            <surname>Brajovic</surname>
          </string-name>
          ,
          <string-name>
            <given-names>B.</given-names>
            <surname>Scalzo</surname>
          </string-name>
          and
          <string-name>
            <given-names>T.</given-names>
            <surname>Constantinides</surname>
          </string-name>
          ,
          <article-title>"Data analytics on graphs. Part I: Graphs and spectra on graphs"</article-title>
          ,
          <source>Foundations and Trends in Machine Learning</source>
          ,
          <year>2020</year>
          , vol.
          <volume>13</volume>
          , no.
          <issue>1</issue>
          , pp.
          <fpage>1</fpage>
          -
          <lpage>157</lpage>
          .
        </mixed-citation>
      </ref>
      <ref id="ref12">
        <mixed-citation>
          [12]
          <string-name>
            <given-names>X.</given-names>
            <surname>Li</surname>
          </string-name>
          ,
          <string-name>
            <given-names>Z.</given-names>
            <surname>Guan</surname>
          </string-name>
          ,
          <string-name>
            <given-names>Q.</given-names>
            <surname>Xu</surname>
          </string-name>
          ,
          <source>Sh. Huang and Ju. Wang, "Graph Classification Model Based on Graph Neural Networks and Graph Distance", 8th International Conference on Computational Science/Intelligence and Applied Informatics (CSII)</source>
          ,
          <year>2021</year>
          , pp.
          <fpage>69</fpage>
          -
          <lpage>74</lpage>
        </mixed-citation>
      </ref>
      <ref id="ref13">
        <mixed-citation>
          [13]
          <string-name>
            <surname>Kuzmenko</surname>
            ,
            <given-names>I.M.</given-names>
          </string-name>
          <article-title>"Graph theory: a study guide for bachelor's degree holders in the specialty 122 "Computer science"</article-title>
          .
          <source>Kyiv: KPI named after Igor Sikorskyi</source>
          ,
          <year>2020</year>
          , 71 p.
        </mixed-citation>
      </ref>
      <ref id="ref14">
        <mixed-citation>
          [14]
          <string-name>
            <surname>Spivak</surname>
            <given-names>I.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Krepych</surname>
            <given-names>S.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Litvynchuk</surname>
            <given-names>M.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Spivak</surname>
            <given-names>S.</given-names>
          </string-name>
          ,
          <article-title>"Validation and data processing in JSON format"</article-title>
          ,
          <source>EUROCON2021 - 19th IEEE International Conference on Smart Technologies, Proceeedings</source>
          ,
          <year>2021</year>
          , pp.
          <fpage>326</fpage>
          -
          <lpage>330</lpage>
        </mixed-citation>
      </ref>
    </ref-list>
  </back>
</article>