Professional Documents
Culture Documents
Download full chapter Recent Developments And The New Direction In Soft Computing Foundations And Applications Selected Papers From The 7Th World Conference On Soft Computing May 29 31 2018 Baku Azerbaijan Shahnaz N Shahb pdf docx
Download full chapter Recent Developments And The New Direction In Soft Computing Foundations And Applications Selected Papers From The 7Th World Conference On Soft Computing May 29 31 2018 Baku Azerbaijan Shahnaz N Shahb pdf docx
https://textbookfull.com/product/soft-computing-applications-
proceedings-of-the-8th-international-workshop-soft-computing-
applications-sofa-2018-vol-i-valentina-emilia-balas/
https://textbookfull.com/product/soft-computing-applications-
proceedings-of-the-8th-international-workshop-soft-computing-
applications-sofa-2018-vol-ii-valentina-emilia-balas/
https://textbookfull.com/product/proceedings-of-the-18th-online-
world-conference-on-soft-computing-in-industrial-applications-
wsc18-bernadetta-kwintiana-ane/
https://textbookfull.com/product/soft-computing-systems-second-
international-conference-icscs-2018-kollam-india-
april-19-20-2018-revised-selected-papers-ivan-zelinka/
https://textbookfull.com/product/proceedings-of-the-tenth-
international-conference-on-soft-computing-and-pattern-
recognition-socpar-2018-ana-maria-madureira/
https://textbookfull.com/product/10th-international-conference-
on-soft-computing-models-in-industrial-and-environmental-
applications-1st-edition-alvaro-herrero/
Shahnaz N. Shahbazova
Janusz Kacprzyk
Valentina Emilia Balas
Vladik Kreinovich Editors
Recent Developments
and the New Direction
in Soft-Computing
Foundations and
Applications
Selected Papers from the 7th World
Conference on Soft Computing,
May 29–31, 2018, Baku, Azerbaijan
Studies in Fuzziness and Soft Computing
Volume 393
Series Editor
Janusz Kacprzyk, Systems Research Institute, Polish Academy of Sciences,
Warsaw, Poland
The series “Studies in Fuzziness and Soft Computing” contains publications on
various topics in the area of soft computing, which include fuzzy sets, rough sets,
neural networks, evolutionary computation, probabilistic and evidential reasoning,
multi-valued logic, and related fields. The publications within “Studies in Fuzziness
and Soft Computing” are primarily monographs and edited volumes. They cover
significant recent developments in the field, both of a foundational and applicable
character. An important feature of the series is its short publication time and
world-wide distribution. This permits a rapid and broad dissemination of research
results.
Indexed by ISI, DBLP and Ulrichs, SCOPUS, Zentralblatt Math, GeoRef, Current
Mathematical Publications, IngentaConnect, MetaPress and Springerlink. The books
of the series are submitted for indexing to Web of Science.
Editors
Recent Developments
and the New Direction
in Soft-Computing
Foundations
and Applications
Selected Papers from the 7th World
Conference on Soft Computing, May 29–31,
2018, Baku, Azerbaijan
123
Editors
Shahnaz N. Shahbazova Janusz Kacprzyk
Department of Information Technology Systems Research Institute
and Programming Polish Academy of Sciences
Azerbaijan Technical University Warsaw, Poland
Baku, Azerbaijan
Vladik Kreinovich
Valentina Emilia Balas Department of Computer Science
Aurel Vlaicu University of Arad University of Texas at El Paso
Arad, Romania El Paso, TX, USA
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Contents
v
vi Contents
Neural Networks
From Traditional Neural Networks to Deep Learning: Towards
Mathematical Foundations of Empirical Successes . . . . . . . . . . . . . . . . . 387
Vladik Kreinovich
Dimensionality Reduction: From Fuzzy Partitions to F-Transforms . . . 399
Irina Perfilieva
Development of Methodological and Functional Principles
of the Intelligent Demographic Forecasting System . . . . . . . . . . . . . . . . 413
Masuma Mammadova and Zarifa Jabrayilova
Correlation Between Low Level Properties of Image and Semantic
Using Algorithm BP Neural Network . . . . . . . . . . . . . . . . . . . . . . . . . . . 423
Guschin Roman, Rozaliev Vladimir, Orlova Yulia,
and Zaboleeva-Zotova Alla
1 Introduction
More and more often social networks, or shall we say data generated by its users,
is an object of research activities being conducted by variety of organization and
S. N. Shahbazova (B)
Azerbaijan Technical University, Baku AZ1073, Azerbaijan
e-mail: shahbazova@gmail.com; shahbazova@berkeley.edu
S. Shahbazzade
The George Washington University, Washington, DC 20052, USA
e-mail: shahbazzade@gwu.edu
The nature of data generated by users of social networks allows us to identify multiple
areas of human life, as well as industrial and corporate activities that could benefit
from analyzing social network data. In the following subsections we describe just a
few of them in the wide category of issues contributing to improvements in: quality
of human life, manufacturing, and corporate visions and goals.
The user’s ability to observe and quickly react to different events—would they be
positive or not—mean that analysis of data can be an important element of sophisti-
cated and intelligent Disaster Management systems. An early detection of disasters,
for example earthquakes or floods, would enable quick interventions that increase
changes of minimizing or even mitigating the effects of disasters. Tools that allow
visualization of social media data would allow for better understanding of dynamics
of investigated calamities. They would allow for close, real-time monitoring of dis-
asters and their progress, and also estimating effectiveness of preventing actions and
provided help. This is closely related to another very important area that could ben-
efit form data analysis: Health Care. In this case, application of big data techniques
Big Data and Fuzzy Based Techniques in Analysis … 5
Facebook [32], Twitter [34], Pinterest [33] and other social sites continue to gen-
erate and provide an uninterrupted stream of data. Development of methods and
approaches that analyze this data from a perspective of the users’ opinions on dif-
ferent products and services will allow companies to identify the customers’ prefer-
ences, as well as their needs and likes. This would lead to a so-called data-driven
manufacturing—a scenario where existing and potential customers influence what
is being manufactured. Analysis of the users’ data would also help identifying weak-
nesses and strengths of manufactured goods. Social media sentiment investigations
could determine if users intent to purchase specific products or if they dislike these
products. All this would provide insight that can be explored and acted upon. Such
analyses could also assess consumers’ interest in a product before it is launched.
Introduction of tools and systems analyzing social network data would affect vari-
ety of service and utility companies. Such systems could change the way companies
provide services to the users, and how these systems could react to changes in the
demand and needs. For example, General Electric (GE) is about to release a system
that uses analysis of social media to support estimation of potential problems in
an electrical grid—it is called Grid IQ [35]. Many aspects of individuals’ lives are
being discussed on the forum of social networks therefore any company that provides
services would benefit a lot from social network data. Another interesting aspect is
related to a so-called geo-tagging where many aspects of collected data is location
sensitive—and this alone will bring valuable information to be analyzed.
6 S. N. Shahbazova and S. Shahbazzade
Understanding of the users’ needs and their attitude to multiple services and products
offered by companies are key elements of building corporate strategies and plans
for future. Also here, vast amounts of data collected during the users’ activities
could provide corporations with valuable information. In such areas like advertising
and marketing any indications regarding the users’ moods, attitudes and opinions
expressed directly or indirectly would be able to change policies and strategies of
companies. The analysis would provide companies with indicators of their social
presence, their ranking, and popularity among users.
One of the most interesting and intriguing aspects of analysis of social network
data could be related to strategies and operations of insurance and financial com-
panies. Information—in a form of opinions, facts or evaluations—that describe the
users’ behaviors, patterns and rules of actions would be of great interest to insur-
ance companies. They could adopt and customize their policies and offer insurance
packages to variety of customers trying to fit their specific needs. Financial compa-
nies could use social media data to improve returns on investment—analysis of the
users’ discussions and posts from the point of few of sentiments and attitudes towards
investment and associated with it expectations can lead to identification of trends,
patterns and motivations of the users. All this would translate into better strategies
and policies of investment polices of companies.
The above section provides a just few examples of areas that can benefit from pro-
cessing, analyzing and modeling social network data. At this stage we would like to
foresee how synthesis of big data methods with fuzzy technologies could contribute
to analysis of social networks.
Most of the work dedicated to analysis of network data is targeting structural
information of networks. The interconnection and relations between actors have been
the main source of information [1, 2, 3, 5, 9, 15]. Some works [4, 6, 7, 14, 19, 23,
27, 28] show an attempt to use additional information that brings different aspects of
analysis. It is well known that each of the nodes/actors as well as connections/relation
is associated with supplementary information. It seems very reasonable that including
that information in analysis procedure and methods would increase the scope of
analysis and allow looking and discovering new insights regarding the users and
their behavior.
A very important aspect of analysis of network data is to look into posts, com-
ments, tweets, or any form of text generated by the users. They contain enormous
amounts of information about users: what is being currently discussed, what are
moods among people, what they like, what types of things invoke positive responses,
what type of things are perceived as negative. It seems critical to be armed with
Big Data and Fuzzy Based Techniques in Analysis … 7
the ability to find answers to these questions—this would result in detecting trends,
popular and/or important topics, determine users’ requirements and expectations
regarding variety of items and events. In such a context, we would like to state that
application of fuzzy-based approaches will enhance the capabilities of data analyt-
ical methods, will enable deeper and more semantic oriented analysis, and what is
also essential will make the obtained outcomes more human-like. Fuzzy methods
targeting summarization of texts [16], sentiment analysis [24, 25], trust inference
and propagation [10, 12], and event detection [18, 30, 31], are just a few examples
of research topics that constitute important issues that can be addressed with, and
will benefit from application of fuzzy-based techniques and methods.
Acknowledgements This scientific work was supported by Grant a Science Development Foun-
dation under the President of the Republic of Azerbaijan, Ministry of Communications and High
Technologies of the Republic of Azerbaijan in order to support projects aimed at the development
of ICT for Grants for 2013 for the 2nd joint ICT Contest (SDF-MCI-MQM-2/ICT-2-2013-7 (13)).
References
1. A.M. Abbasov, Information boom: new trends and expectations, in Soft Computing: State of
the Art Theory and Novel Applications (Studies in Fuzziness and Soft Computing), ed. by R.R.
Yager, A.M. Abbasov, M.Z. Reformat, S.N. Shahbazova (Springer, Berlin, 2012), pp. 1–12
2. S.P. Borgatti, Centrality and network flow. Soc. Netw. 27(1), 55–71 (2005)
3. J. Boyd, M. Everett, Relations, residuals, regular interiors, and relative regular equivalence.
Soc. Netw. 21(2), 147–165 (1999)
4. M. Brunelli, M. Fedrizzi, M. Fedrizzi, Fuzzy m-ary adjacency relations in social network
analysis: Optimization and consensus evaluation. Inform. Fusion 17, 36–45 (2014)
5. Y. Cao, J. Cao, M. Li, Distributed data distribution mechanism in social network based on fuzzy
clustering, in Foundations and Applications of Intelligent Systems, Advances in Intelligent
Systems and Computing, vol. 213, ed. by F. Sun, et al. (Springer, Berlin Heidelberg, 2014),
pp. 603–620
6. T. Casasús-Estellés, R.R. Yager, Fuzzy concepts in small worlds and the identification of
leaders in social networks, in IPMU 2014, Part II, CCIS 443, ed. by A. Laurent, et al. (Springer
International Publishing, Switzerland, 2014), pp. 37–45
7. S. Elkosantini, D. Gien, A dynamic model for the behavior of an operator in a company, in
Proceedings of the 12th IFAC Symposium on Information Control Problems in Manufacturing,
France vol. 2, (2006), pp. 187–192
8. S. Elkosantini, D. Gien, Human behavior and social network simulation: fuzzy sets/logic
and agents-based approach, in Proceedings of the 2007 Spring Simulation Multi-conference
SpringSim ‘07, vol. 1 (2007), pp. 102–109
9. L. Freeman, The Development of Social Network Analysis (Empirical Press, Vancouver, 2006)
10. F. Hao, G. Min, M. Lin, C. Luo, L.T. Yang, IEEEMobiFuzzyTrust: an efficient fuzzy trust
inference mechanism in mobile social networks. IEEE Trans. Parallel and Distrib. Syst. 25(11),
2944–2955 (2014)
11. R. Hannemanand, M. Riddle, Introduction to Social Network Methods (University of California,
Riverside, 2005)
12. S. Kim, S. Han, The method of inferring trust in web-based social network using fuzzy logic, in
Proceedings of the International Workshop on Machine Intelligence Research (2009) pp. 140–
144
8 S. N. Shahbazova and S. Shahbazzade
13. G.J. Klir, B. Yuan, Fuzzy Sets and Fuzzy Logic: Theory and Applications (Pretience Hall, 1995)
14. P.N. Krivitsky, M.S. Handcock, A.E. Raftery, P.D. Hoff, Representing degree distributions,
clustering, and homophily in social networks with latent cluster random effects models. Soc.
Netw. 31, 204–213 (2009)
15. M.J. Lanham, G.P. Morgan, K.M. Carley, Social network modeling and agent-based simulation
in support of crisis de-escalation. IEEE Trans. Syst. Man Cybern. 44(1), 103–110 (2014)
16. J. Liu, Fuzzy modularity and fuzzy community structure in networks. Eur. Phys. J. B 77(4),
547–557 (2010)
17. X.H. Liu, Y.T. Li, F.R. Wei, M. Zhou, Graph-based multi-tweet summarization using social
signals, in Proceedings of COLING 2012 (2012), pp. 1699–1714
18. T. Matuszka, Z. Vincellér, S. Laki, On a keyword-lifecycle model for real-time event detection
in social network data, in Proceedings of 4th IEEE International Conference on Cognitive
Infocommunications (2013), pp. 453–458
19. W. Pedrycz, S.-M. Chen (eds.), Social Networks: A Framework of Computational Intelligence,
Studies in Computational Intelligence, vol. 526 (Springer International Publishing Switzerland,
2014)
20. M. Z. Reformat, R. R. Yager, Using tagging in social networks to find groups of com-
patible users, in Proceedings of Join IFSA World Congress and NAFIPS Annual Meeting
(IFSA/NAFIPS), Edmonton, Canada (2013), pp. 697–702
21. S.N. Shahbazova, Application of fuzzy sets for control of student knowledge//applied and
computational mathematics. Int. J. 10(1), 195–208 (2011). Special Issue on Fuzzy Set Theory
and Applications
22. J. Scott, Social Network Analysis. A Handbook (Sage, London, 2000)
23. T.A.B. Snijders, C. Baerveldt, A multilevel network study of the effects of delinquent behavior
on friendship evolution. J. Math. Sociol. 27, 123–151 (2003)
24. T.A.B. Snijders, Statistical models for social networks. Annu. Rev. Soc. (2011)
25. G. Stakias, M. Psoras, and M. Glykas, Fuzzy cognitive maps in social and business network anal-
ysis, in Business Process Management, SCI 444 ed. by M. Glykas (Springer, Berlin Heidelberg,
2013), pp. 241–279
26. D.N. Trung, J.J. Jung, L.A. Vu, A. Kiss, Towards modeling fuzzy propagation for sentiment
analysis in online social networks: a case study on TweetScope, in Proceedings of 4th IEEE
International Conference on Cognitive Infocommunication (2013), pp. 331–337
27. D.N. Trung, J.J. Jung, Sentiment analysis based on fuzzy propagation in online social networks:
a case study on TweetScop. Comput. Sci. Inf. Syst. 11(1), 215–228 (2014)
28. F. Vega-Redondo, Complex Social Networks (Cambridge University Press, 2007)
29. R.R. Yager, Intelligent social network analysis using granular computing. Int. J. Intell. Syst.
23, 1196–1219 (2008)
30. R.R. Yager, M.Z. Reformat, Looking for like-minded individuals in social networks using
tagging and fuzzy sets. IEEE Trans. Fuzzy Sets 21(4), 672–687 (2013)
31. L.A. Zadeh, Fuzzy sets. Inf. Control 8, 338–353 (1965)
32. https://www.facebook.com. Accessed 7 Nov 2015
33. https://www.pinterest.com. Accessed 7 Nov 2015
34. https://twitter.com. Accessed 7 Nov 2015
35. http://www.gegridsolutions.com/demandopt/Catalog/GridIQ.htm. Accessed 7 Nov 2015
Application of Fuzzy Pattern Recognition
in the Recruitment of IT Specialists
Abstract The article the task of the management of the IT specialists market reduced
to decision-making on selecting the policy for the coordination of supply and demand.
The fuzzy situational models for supply and demand are proposed. Decision support
methods for the recruitment of IT specialists based on fuzzy pattern recognition
taking into account the multivariate nature of the coordination of supply and demand
is developed.
1 Introduction
The specific features of the IT labor market, the uncertainty of information flows
about its state, the much-varied profile of data on IT specialists and the difficulties
of measuring it, and the ambiguity of indicators characterizing it predetermine the
variety of possible fuzzy states of supply and demand for IT specialists and the multi-
option pattern of reconciling them. Comparisons and evaluations of these conditions
and the choice of the reconciliation policy can be efficiently accomplished with the
use of intellectual methods and technologies [6–8]. The latter help to integrate the
versatile information on the supply and demand conditions and facilitate the devel-
opment of a set of alternatives for managerial decisions and the selection of the most
efficient one among them. Management in such systems is considered as a process
of identification of the demand and supply condition at a given point in time and
decision making suitable in the current situation.
The listed features identify the problem of management of the labor market of IT
specialists as semi-structured and difficultly formalized tasks [9–11]. Obviously, for
the solution of such problems, the use of intelligent methods and technologies is the
most effective. In this interpretation, solution of the management problem of the IT
specialists market assumes:
• development of situational models of supply and demand for IT specialists;
• comparison and evaluation of fuzzy situation of supply and demand in the current
situation;
• deciding on the choice of a policy for their harmonization [4].
For the most flexible expression of continuously changing management situa-
tions and the correlation of supply and demand, we use the model of fuzzy pattern
recognition.
The study develops decision support methods for the recruitment of IT special-
ist on the multivariate coordination of supply and demand based on fuzzy pattern
recognition.
Assume are two sets of fuzzy situations describing the conditions of demand Ṽk and
supply S̃q in the IT specialist labor market [4]:
Ṽi = < μl j (Vi ) >, < μcr (Vi ) >, < μu z (Vi ) > = μVi (y) y
S̃g = < μl j Sg >, < μcr Sg >, < μu z Sg > = μ Sg (y) y
Here:
• μl j (Vi ) : V × L → [0, 1]—the degree of conformity of the vacancy V i by the indi-
cators L and L = {l1 , l2 , . . . , ln }—is the set of personal features (characteristics)
required in a candidate to a particular position (job, workplace);
Application of Fuzzy Pattern Recognition in the Recruitment … 11
3 Problem Solution
For the intelligent management of supply and demand in the IT specialists labor
market, the problem of the supply and demand conformity decision can be posed as
a task of fuzzy pattern recognition. The problem of pattern recognition is based on
a fuzzy situational analysis and determination of the degree of similarity of fuzzy
situations.
The procedure of pattern recognition involves the following steps:
• real situations are determined (search patterns of IT specialists) in accordance
with the indicator’s values that characterize each applicant for the announced
vacancy;
• reference situations are determined (search patterns of requests) in accordance
with the indicators’ values that characterize the employer’s requirements to the
applicant for the vacancy;
• calculate the degree of fuzzy similarity of the reference situation to each real situ-
ation in accordance with the chosen measure of assessment of similarity between
two fuzzy situations;
• real situation is revealed, which has the greatest proximity rate with the refer-
ence situation. In other words, a decision is made to hire a IT specialist (supply)
12 M. Mammadova and Z. Jabrayilova
for the announced vacancy (demand), which most closely corresponds to the
requirements of the employer.
The works [10, 12] review various measures to determine proximity rate between
two fuzzy situations, involving single-stage or multistage measurement procedures.
The case study uses the degree fuzzy inclusion of situation S̃g in fuzzy situation
Ṽi , and the degree of fuzzy equality of S̃g and Ṽi as the measure of identifying the
proximity rate of fuzzy real and reference situations [12].
In the work [4], the authors identified possible scenarios for coordinating supply and
demand for IT specialists during the recruitment. Hence, after the completion of the
process of recognizing the most acceptable pair of “fuzzy reference image—fuzzy
real image” in terms of proximity rate, there may be several possible scenarios among
the sets of real search patterns of IT specialists (supply) and reference search patterns
of the query (demand):
Scenario 1
One vacancy (employer’s request)—one applicant (IT specialist), i.e. “one fuzzy
reference image—one fuzzy real image”.
Method for decision making. In this case, if the degree of fuzzy similarity of the
two situations (i.e. S̃g and Ṽi ) is not less than the threshold accepted by the employer
(for example, ψ∈ [0,6; 1]), then a hiring decision is made.
The degree of fuzzy inclusion of a fuzzy situation S̃g in a situation Ṽi is determined
based on the following formula:
The situation S̃g is considered fuzzy inclusive in the situation Ṽi (..⊂Ṽi ), if the
∼
degree of inclusion of S̃g in a situation Ṽi is not less than the inclusion threshold (ψ∈
[0,6; 1] determined by the management conditions), i.e. μ ( S̃g , Ṽi ) ≥ ψ.
The degree of fuzzy equality Ṽi and S̃g is defined by the following formula:
μ( S̃g , Ṽi ) = ∨( S̃g , Ṽi )& ∨ (Ṽi , S̃g ) = & μ(μ Sg (y), μVi (y)
= min min(max(1 − μ Sg (y), μVi (y)), max(1 − μVi (y), μ Sg (y))) (2)
y∈Y
Application of Fuzzy Pattern Recognition in the Recruitment … 13
Situations S̃g and Ṽi are considered to be fuzzy equal S̃g ≈ Ṽi , if μ( S̃g , Ṽi ) ≥
ψ, ψ ∈ [0, 7; 1], where ψ is a certain threshold of fuzzy equality of situations.
Scenario 2
A few applicants (IT specialists) are eligible to the employer in accordance with the
accepted measure of the similarity of the two fuzzy situations, i.e. “one fuzzy refer-
ence image—more fuzzy real image”. The applicants form subset of fuzzy situations
(alternatives), the one of which must be chosen that corresponds to the most suitable
candidate.
Therefore, there is a set μ( S̃g , Ṽi ), g = 1, η, η ≤ q) , the elements of which
satisfy the following condition: (μ( S̃g , Ṽi ), g = 1, η ) ≥ ψ.
In this case, the employer, acting as an expert (DM), can be offered various
methods for decision making:
Scenario 2.1. The task of decision making is reduced to the comparison of simi-
larity of reference and real situations by the degree of possessing the criteria charac-
terizing the applicants for the vacancy. The best alternative (applicant) is that, who
has the greatest proximity rate by coincidence of criteria and the degree of their
possession.
Method for decision making. In this case, the sought is the real situation, which
provides the maximum value in terms of the degree of reference similarity. I.e.:
the specific IT specialists determine the degree of fuzzy similarity of fuzzy real
situations with the reference situation by the following step [19]:
(a) Based on the “convolution” μlg j S̃g , Ṽi , j = 1, n, the degree of fuzzy
similarity between real and reference situations is determined by personal
characteristics (L):
n
μ L S̃g , Ṽi = w j μl j S̃g , Ṽi .
j=1
(b) Based on the “convolution” μcg f S̃g , Ṽi , f = 1, m—the degree of fuzzy
similarity between real and reference situations is determined in terms of
competences (C):
m
μC S̃g , Ṽi = w f μc f S̃g , Ṽi .
f =1
p
μU S̃g , Ṽi = w γ μu γ S̃g , Ṽi .
γ =1
(d) Based on the obtained results and coefficients of relative importance L, C and
U—wL , wC , wU , the similarity rates of the real situation with the reference ones
are determined:
The selected fuzzy real situation corresponds to the search pattern of the applicant,
who has the greatest degree of fuzzy similarity with the reference pattern of the
vacancy, and is accepted as the best solution.
Scenario 2.3. The list of evaluation criteria is extended, and the input situations
are re-defined (re-examined) and the recognition procedures are repeated.
Scenario 3
Several employers are identified, who are interested in hiring one IT specialist, i.e.
“more fuzzy reference image—one fuzzy real image”.
In this case, there is an inverse task: a subset of fuzzy reference situations (alter-
natives) is presented, which is represented by the vacancies with the corresponding
hiring conditions, among which a IT specialist has to make a choice according to
his/her preferences. According to the preferences of the latter, there can be the variants
of scenario 3 listed below.
Scenario 3.1. The proximity rate of the claims of the IT specialist with the criteria
characterizing the conditions of employment shall be compared, and decision shall
be made on the greatest coincidence of the degree of possession of the criteria.
Method for decision making. The task of decision making is reduced to the compar-
ison of the similarities between real and standard situations in terms of the similarity
rate of the conditions offered by employers and the claims of the applicant. The best
vacancy has the greatest degree of similarity by the coincidence of the criteria that
characterize the conditions offered by employers, and the applicant’s claims. Thus,
assume that:
In this case, the pair with the maximum value of the degree of fuzzy similarity
situations is defined by the following formula:
The fuzzy reference situation is accepted as the best solution, which is corre-
sponding to the search pattern of the vacancy that has the greatest degree of fuzzy
similarity with the applicant’s real pattern.
Scenario 3.2. The task of decisions making shall be reduced to the multi-objective
task of choosing the best solution, taking into account the relative importance of the
criteria that characterize workplaces.
Method for decision making. The task of decisions making shall be reduced to the
multi-objective task of choosing the best solution, taking into account the relative
importance of the indicators characterizing the preferences of the IT specialist (U),
expressed in terms of his/her requirements for a IT profile vacancy.
If ωγ , γ = 1, p—the coefficients of the relative importance of the indicators
characterizing the criterion U, then a fuzzy reference situation that has the greatest
16 M. Mammadova and Z. Jabrayilova
degree of fuzzy similarity with the applicant’s real image is determined based on the
following formula:
⎧ ⎫
⎨ p ⎬
μ( S̃d , Ṽz )∗ = max wγ · μu γ ( S̃d , Ṽz ), z = 1, f .
⎩ ⎭
γ =1
Table 3 Fuzzy degrees of matching the indicators, which are characterizing the vacancy V1 , by
the applicants
L, C, U S1 S2 S3 S4
l1 0.97 0.75 0.78 0.97
l2 0.9 0.84 0.82 0.90
l3 0.92 0.86 0.72 0.88
l4 0.9 0.77 0.9 0.85
c1 0.94 0.98 0.92 0.93
c2 0.79 0.93 0.78 0.7
c3 0.77 0.93 0.65 0.8
u1 0.98 0.90 0.95 0.95
u2 0.97 0.65 0.7 0.95
u3 0.95 0.49 0.97 0.9
0.97 l1 ; 0.9/l2 ; 0.92/l3 ; 0.9/l4 ; 0.94/c1 ; 0.79/c2 ;
S̃1 =
0.77/c3 ; 0.98/u 1 ; 0.97/u 2 ; 0.95/u 3
0.75 l1 ; 0.84/l2 ; 0.86/l3 ; 0.77/l4 ; 0.98/c1 ;
S̃2 =
0.93/c2 ; 0.93/c3 ; 0.9/u 1 ; 0.65/u 2 ; 0.69/u 3
0.78 l1 ; 0.82/l2 ; 0.72/l3 ; 0.9/l4 ; 0.92/c1 ; 0.78/c2 ;
S̃3 =
0.65/c3 ; 0.95/u 1 ; 0.7/u 2 ; 0.97/u 3
0.97 l1 ; 0.9/l2 ; 0.88/l3 ; 0.85/l4 ; 0.93/c1 ;
S̃4 =
0.7/c2 ; 0.8/c3 ; 0.95/u 1 ; 0.95/u 2 ; 0.9/u 3
Stage 4: Using the formula (2) fuzzy equality degrees of the reference Ṽ1 and real
situations S̃1 , S̃2 , S̃3 , S̃4 are determined: (a) by personal indicators (L); (b) in terms
of competences (C); and (c) from the prism of the requirements for the vacancy (U):
Determination of the fuzzy equality degree of the reference Ṽ1 and real situations
S̃1 , S̃2 , S̃3 , S̃4 , according to the personal indicators (L):
μ L (Ṽ1 , S̃1 ) = min(max(1 − 0.97, 0.97), max(0.97, 1 − 0.97))& min(max(1 − 0.98, 0.9),
max(0.96, 1 − 0.9))&& min(max(1 − 0.99, 0.92), max(0.99, 1 − 0.92))
& min(max(1 − 0.96, 0.9), max(0.96, 1 − 0.9)) = min(0.97, 0.97)& min(0.9, 0.98)
& min(0.92, 0.99)& min(0.9, 0.96) = 0.97&0.9&0.92&0.9 = 0.9
Determination of the fuzzy equality (similarity) degree of the applicants with the
vacancy requirements V 1 in terms of their competencies (C):
In accordance with the obtained results taking into account all indicators, fuzzy
equality degree φ(Ṽ1 , S̃g ), g = 1, 4 of the reference V 1 and search images of the
real situations S̃1 , S̃2 , S̃3 , S̃4 is determined:
μ(Ṽ1 , S̃1 ) = μ L (Ṽ1 , S̃1 )&μC (Ṽ1 , S̃1 )&μU (Ṽ1 , S̃1 ) = 0.9&0.77&0.95 = 0.77
μ(Ṽ1 , S̃2 ) = μ L (Ṽ1 , S̃2 )&μC (Ṽ1 , S̃2 )&μU (Ṽ1 , S̃2 ) = 0.75&0.93&0.49 = 0.49
μ(Ṽ1 , S̃3 ) = μ L (Ṽ1 , S̃3 )&μC (Ṽ1 , S̃3 )&μU (Ṽ1 , S̃3 ) = 0.72&0.65&0.7 = 0.65
μ(Ṽ1 , S̃4 ) = μ L (Ṽ1 , S̃4 )&μC (Ṽ1 , S̃4 )&μU (Ṽ1 , S̃4 ) = 0.85&0.7&0.9 = 0.7