=Paper=
{{Paper
|id=Vol-2914/Paper2.pdf
|storemode=property
|title=The Task of Assessing the Effectiveness of University Employees in Fuzzy Decisions
|pdfUrl=https://ceur-ws.org/Vol-2914/paper2.pdf
|volume=Vol-2914
|authors=Natalia V. Apatova,Irina N. Ostapenko,Roman S. Usenko
}}
==The Task of Assessing the Effectiveness of University Employees in Fuzzy Decisions==
https://ceur-ws.org/Vol-2914/paper2.pdf
13
The Task of Assessing the Effectiveness of University
Employees in Fuzzy Decisions*
Natalia V. Apatova 1[0000-0003-4066-3821], Irina N. Ostapenko 1[0000-0003-1946-7586]
and Roman S. Usenko 1[0000-0003-0095-1183]
1V.I. Vernadsky Crimean Federal University
Russian Federation, Republic of Crimea, Simferopol
i.n.ostapenko@mail.ru
Abstract. Talented employees, their intelligence, creativity, and ability to cre-
ate something new are some of the main competitive advantages that determine
the success of an organization's development. Due to dynamic changes in the
labor market, the expectations and requirements of companies for employees
are changing. Human skills and abilities have come to the fore in social produc-
tion. Any manager, based on certain indicators, can evaluate his subordinates
and their ability to quickly and efficiently complete the task. Quite often, quali-
tative indicators assessed by experts are used to assess the effectiveness of ac-
tivities. The article discusses the possibilities of using the apparatus of fuzzy
mathematics to assess the effectiveness of university employees. The study de-
scribes a method of a quantitative assessment of their effectiveness, which al-
lows for competent management of the university. The main advantage of fuzzy
models, in comparison with mathematical models based on classical mathemat-
ical tools, is associated with the possibility of using significantly smaller
amounts of input data about the system for their development. In this case, the
input data can be approximate, indistinct. The theory of fuzzy sets allows you to
formally describe non-strict fuzzy concepts and provides an opportunity to un-
derstand the processes occurring in conditions of a high degree of uncertainty.
Keywords: Educational Organization of Higher Education, Employee Perfor-
mance Evaluation, Fuzzy Mathematics, Membership Function.
1 Introduction
The modern economy is characterized by an extremely fast pace of changes in the
business environment caused by technological innovations, the intensive development
of new industries and activities, changes in consumer needs, and increased competi-
tion. In these conditions, the role of personnel increases, their ability to develop their
*
Copyright 2021 for this paper by its authors. Use permitted under Creative Commons License
Attribution 4.0 International (CC BY 4.0).
14
labor potential, and use potential opportunities in the world of work to achieve the
goals of the enterprise.
Talented employees, their intelligence, creativity, and ability to create something
new are some of the main competitive advantages that determine the success of an
organization's development. Due to the dynamic changes in the labor market, the
expectations and requirements of companies for employees are changing. Human
skills and abilities came to the fore in social production.
Any manager, based on certain indicators, can evaluate his subordinates and their
ability to quickly and efficiently perform the assigned task. Quite often, to assess the
effectiveness of activities, qualitative indicators are used, which are expert assess-
ments.
The purpose of the article is to consider the methodology for the quantitative as-
sessment of the effectiveness of university employees, based on the use of the appa-
ratus of fuzzy mathematics.
2 Main part
The problems of human capital development and the assessment of the effectiveness
of investments in human capital, in particular in education, were studied by such sci-
entists: G.S. Becker [1], I. Šlaus and G. Jacobs [2], J. Mincer [3], R.J. Barro [4], V.M.
Porokhnya [5], N.R. Kelchevskaya, E.V. Shirinkina [6], and others.
Recently, fuzzy logic methods have been widely used to assess the effectiveness of
employees [7-11]. The methodology proposed in the article helps to assess the effec-
tiveness of employees using these methods. The main advantage of fuzzy models
proposed for use, in comparison with traditional mathematical models, is associated
with the possibility of using much smaller amounts of input data about the system for
their development. In this case, the input data can be approximate, indistinct. The
theory of fuzzy makes it possible to formally describe non-strict fuzzy concepts and
provides an opportunity to understand the processes taking place in conditions of a
high degree of uncertainty.
A fuzzy set is a set of elements of an arbitrary nature, concerning which it is im-
possible to assert with complete certainty whether this or that element of the set under
consideration belongs to this set or not. Fuzzy models are based on a system of rules,
which is usually formed based on expert knowledge about the object of research. This
approach is called knowledge acquisition and is effective if the expert has full
knowledge of the system and can express this knowledge in verbal form and convey
it.
The ability to draw on expert knowledge in this procedure is critical to success.
However, in the case of measuring the level of creativity, the expert's knowledge is
often incomplete, inaccurate, poorly formulated, and may even contain contradictions.
Also, this knowledge is subjective, that is, the opinions of individual people about the
functioning of the same information system may differ. Following the methodology
for assessing the indicator of interest to us, using fuzzy modeling, it is possible to
develop an expert system. At the output of the expert system, according to the input
15
data, an estimate of the indicator of interest to us will be obtained based on the criteria
that determine it.
Let us consider the issue of modeling the effectiveness of an employee's activity,
more precisely, the effectiveness of the employee's return, to the educational organi-
zation of higher education using an expert system developed using the software tools
of the MATLAB package. By the described approach, it is necessary to develop an
expert system, which will have to make it possible to assess the effectiveness of an
employee of the university based on the given input variables and their subjective
assessments.
Following this approach to assessing the effectiveness of an employee using fuzzy
modeling, it will be necessary to develop an expert system. It will have to allow eval-
uating the employee's performance based on the given input variables and their sub-
jective assessments.
When assessing the effectiveness of an employee, the following factors can be de-
termined that affect the parameter being assessed:
─ Professional competence level (F1)
─ Self-education (F2)
─ Employee age (F3)
These factors can be used as input variables, and their level can be set expertly
based on the results of testing according to the method presented in Table 1.
Table 1. Levels of input variables and types of term sets.
Input variables Type of term set
Professional competence level (F1) Very low (VL)
Low (L)
Medium (M)
Above average (AA)
High (H))
Self-education (F2) Occasionally (OS)
Regularly (R)
Employee age (F3) Young (YN)
Medium (M)
Mature (F)
The input variable "Professional competence level (F1)" should be presented in points
on a scale from 0 to 100 and can be represented as 5 term sets with the following
gradation:
─ Very low (0-20);
─ Low (20-40);
─ Medium (40-60);
─ Above average (60-80);
─ High (80-100).
16
For the problem being solved, for the input variable "The level of professional
competencies (F1)", it is necessary to choose a trapezoidal membership function,
which is a generalization of the triangular one and allows you to determine the kernel
of a fuzzy set in the form of an interval. The specified parameters of the input variable
"Level of professional competence (F1)" in the editor of membership functions are
shown in Figure 1.
The input variable "Self-education (F2)" (which includes work on their level of
scientific and educational qualifications, including official refresher courses) should
be represented as 2 term sets with the following gradation:
─ Episodic (1)
─ Regular (2).
Fig. 1. Parameters of the input variable "Level of professional competence (F1)" in the editor of
membership functions.
For the problem being solved for the input variable "Self-formation (F2)" it is possi-
ble to choose a sigmoid membership function. This type of function allows you to
form membership functions for which the values starting from some value of the ar-
gument and up to + (-) infinity are equal to 1.
These functions are useful for specifying linguistic terms such as “low” or “high”.
The specified parameters of the input variable "Ca-formation (F2)" in the editor of
membership functions are shown in Figure 2.
And, finally, for the problem to be solved for the input variable "Employee age
(F3)", it is also necessary to choose a trapezoidal membership function with 3 term
sets:
17
─ Young (25-35);
─ Medium (35-50);
─ Mature (over 50; for example, 50-80).
Fig. 2. Parameters of the input variable "Self-formation (F2)" in the editor of membership func-
tions.
The specified parameters of the input variable "Employee age (F3)" in the editor of
membership functions are shown in Figure 3. At the output of the expert system,
based on the input data, one can obtain, for example, an assessment of the employer's
investment in an employee in terms of efficiency. In our case, we will choose Y - the
level of "efficiency of return" of the employee from 0 to 1, so that in the future it will
be convenient to compare with the result of logistic regression (Table 2).
Table 2. Output variable levels and types of term sets.
Output variable Term set type
Efficiency performance (Y) Very low (VL)
Low (L)
Average (AV)
Above average (AA)
High (H)
Very high (VH)
18
The specified parameters of the output variable "Employee efficiency (Y)" in the
editor of membership functions are shown in Figure 4.
Fig. 3. Parameters of the input variable "Employee age (F3)" in the editor of membership func-
tions.
Fig. 4. Parameters of the output variable "Employee efficiency (Y)" in the editor of member-
ship functions.
19
After determining the types of membership functions for input and output variables
and their terms for the developed expert system, it is necessary to set the rules of
fuzzy inference.
The level of influence of input variables on the output is described in Table 3.
Table 3. Level of Influence of input variables.
Input Variables Y
Professional competence level (F1) High
Self-education (F2) Medium
Employee age (F3) Low
To understand how the input variables affect the output, a matrix of fuzzy inference
rules should be drawn up. Since we have three input variables with different numbers
of gradations, the number of fuzzy inference rules determine by multiplying the num-
ber of gradations of input variables, which in our case will be 5 * 2 * 3 = 30. The
compilation of fuzzy inference rules by the developed rule base is shown in Figure 5.
As a result of processing the values of the input variables, after the formation of the
output fuzzy set and its subsequent defuzzification, a clear value of the output varia-
ble will be found.
Fig. 5. Drawing up rules for fuzzy inference.
So, suppose that in the course of the expert assessment, the values of three input vari-
ables were obtained. These values of the input variables can be set in the Rule Viewer
window, and at the same time, the value of the output variable can be obtained at the
output. In Figure 6, as an example, the value of the output variable "Employee effi-
ciency (Y)" is found for the following initial values of the input variables:
─ The level of professional competencies (F1) - 45;
20
─ Self-education (F2) - 1;
─ The employee's age (F3) is 52.
Fig. 6. The result of modeling a fuzzy inference system for given values of input variables
For the given values of the input variables, the expert system evaluates the employee
at 0.254 points out of 1.
Similarly, it is possible to obtain the numerical characteristics of the efficiency for
all employees, which are presented in Table 4. As can be seen from Table 5, the de-
veloped expert system characterizes the employee's efficiency in the range from 0 to
1, while the minimum value is 0.185 and the maximum value is 0.921.
The graphical interface of the MATLAB package allows you to get a graph of the
dependence of the output variable on the values of any of the input variables. Figure 7
shows the dependence of the output variable "Employee efficiency (Y)" on the input
variable "Level of professional competencies (F1)".
Also, the graphical interface of the MATLAB package allows you to get the sur-
face of the dependence of the output variable when changing two input variables with
a fixed value of the third variable. Figure 8 shows the dependence of the output varia-
ble "Employee efficiency (Y)" on the level of professional competencies and self-
education, at a fixed age of the employee.
21
Table 4. Assessment of employee performance
№ F1 F2 F3 Y
1 45 1 52 0,254
2 62 2 32 0,699
3 52 2 36 0,570
4 92 2 34 0,912
5 78 2 41 0,784
6 30 1 50 0,185
7 48 1 60 0,251
8 74 2 67 0,749
9 64 1 38 0,507
10 56 2 56 0,478
11 75 1 57 0,577
12 36 1 48 0,218
13 45 2 28 0,571
14 98 2 37 0,914
15 42 1 36 0,254
16 30 2 44 0,250
17 58 2 46 0,620
18 86 2 62 0,921
19 35 2 42 0,250
20 54 1 47 0,253
Fig. 7. Dependence of the output variable "Employee efficiency (Y)" on the input variable
"Level of professional competencies (F1)".
22
Fig. 8. The surface of the dependence of an output variable on two input variables.
The presented graphs allow us to say that the dependence of the employee's efficiency
on the factors determining it, following the compiled rules of fuzzy inference, can be
both linear and non-linear.
Having considered the results of Table 4, we can conclude that the effectiveness of
an employee from the factors determining it can be obtained numerically using an
expert system built on fuzzy logic. However, the described approach can be rather
complicated for specialists who do not know how to use fuzzy logic tools.
The simplest way to assess the effectiveness of an employee from the factors that
determine it can be the use of a linear regression equation.
For the employee efficiency and input variables calculated using an expert system
based on fuzzy logic, the following multiple regression equation can be obtained:
Y 0.2876 0.0101F1 0.1769 F2 0.0015 F3
The resulting equation allows you to calculate the employee's efficiency by speci-
fying the values of the input variables (see Table 5).
Table 5 Regression statistics equation
Feature Value
Multiple R 0,972
R-square 0,945
Normalized R-square 0,935
Standard error 0,066
Observations 20
23
Table 5 ("Regression statistics") and table 6 ("Analysis of variance") suggest that
the obtained regression equation has fairly high accuracy (R2 = 0.94) and is statisti-
cally significant.
Table 6. ANOVA table
df SS MS F Significance F
Regression 3 1,20714 0,40238 92,3174 2,58E-10
The remainder 16 0,069739 0,004359
Total 19 1,276879
To obtain more accurate results, you can use nonlinear modeling methods based, for
example, on the use of neural networks [12].
3 Conclusion
In the context of informatization and digitalization, the value of human capital has
increased as a factor in the company's success. The development of new knowledge
and the adoption of managerial decisions for business success is the merit of the per-
son. Today, the latest technologies provide a competitive advantage for the organiza-
tion, but under equal technological conditions, talented employees, their potential, and
knowledge are the key to the competitive advantages of the organization.
The approach using the apparatus of fuzzy mathematics in situations of a high de-
gree of uncertainty allows you to operate with high-quality input data. A quantitative
assessment of the personality quality under consideration can be useful in assessing
the talent, creativity of an individual, and the creative potential of an organization.
Assessment of the level of creativity of an individual is the most important stage in
assessing the level of talent, the usefulness of an employee for the organization.
The proposed assessment cannot very accurately reflect the phenomenon so diffi-
cult to quantify, but it can be useful for monitoring the development of the creative
component of intellectual capital. The presented results allow us to say that the de-
pendence of the employee's efficiency on the factors determining it can be obtained
numerically using an expert system built on fuzzy logic.
References
1. Becker, G.S.: Human Capital. A Theoretical and Empirical Analysis, with Special Refer-
ence to Education. Chicago, University of Chicago Press, 1993.
2. Šlaus, I. and Jacobs, G.: Human Capital and Sustainability. Sustainability, 3(1), 2011, pp.
97–154
3. Mincer, J.: Human Capital and the Labor Market. A Review of Current Research. Educa-
tional Researcher, 1989, vol. 18, no.4, pp.27–34.
4. Barro, R.J.: Education and Economic Growth. Annals of Economics and Finance, 2013,
vol. 14, P. no.2, pp. 301–328.
24
5. Porokhnya, V.M.: Intelektualʹnyy kapital ekonomichnoho zrostannya: navch. posibn. [In-
tellectual Capital of Economic Growth: textbook.]. Zaporizhzhya: KPU, 2012. (in Ukraini-
an)
6. Kelchevskaya, N. R. and Shirinkina, E. V.: Regional Determinants of Effective Use of
Human Capital in the Digital Economy. Ekonomika regiona [Economy of Region]. 2019,
no.15(2), pp. 465-482 (in Russian)
7. Ostapenko, I.N., Usenko, R.S.: Modelirovaniye pokazatelya urovnya tvorcheskogo razviti-
ya lichnosti [Modeling an Indicator of the Level of Creative Development of a Person].
Digital economy: information technology and models: monograph. 2018, pp.258-282. (in
Russian)
8. Jevšček, M.: Competencies Assessment Using Fuzzy Logic. Journal of Universal Excel-
lence. 2016, no.2, p. 187-202.
9. Houe, R., Grabot, B., Tchuente, G.: Fuzzy Logic In Competence Management. European
Society for Fuzzy Logic and Technology: 7th conference. 2011, p.651-656.
10. Macwan, N., Srinivas, S.: Performance Appraisal Using Fuzzy Evaluation Methodology.
International Journal of Engineering and Innovative Technology. 2013, no.3, p. 324-329.
11. Krichevskiy, M.L., Martynova, Yu.A.: Instrumenty iskusstvennogo intellekta pri otsenke
effektivnosti investitsionnogo proekta [Instruments of Artificial Intelligence in Assessment
of Effectiveness of Investment Project]. Kreativnaya ekonomika. 2018, no.12, pp.1105-
1118. doi: 10.18334/ce.12.8.39265. (in Russian)
12. Apatova, N.V., Ostapenko, I.N., Usenko, R.S.: Decision-Making on Human Individual
Capital Investment. CEUR Workshop Proceedings. 2020.