=Paper=
{{Paper
|id=Vol-2122/paper-2
|storemode=property
|title=Bank Attractiveness Evaluation Method Based on Soft
Computing in the Analytic Hierarchy Process
|pdfUrl=https://ceur-ws.org/Vol-2122/paper_127.pdf
|volume=Vol-2122
|authors=Davyd Dabahian,Mykhailo Godlevskyi
|dblpUrl=https://dblp.org/rec/conf/icteri/DabahianG18
}}
==Bank Attractiveness Evaluation Method Based on Soft
Computing in the Analytic Hierarchy Process==
https://ceur-ws.org/Vol-2122/paper_127.pdf
Bank attractiveness evaluation method based on soft
computing in the analytic hierarchy process
Davyd Dabahian and Mykhailo Godlevskyi
National Technical University Kharkiv Polytechnic Institute, Kyrpychova str.,21,
Kharkiv, Ukraine 61002
{d.dabagyan@gmail.com, god_asu@kpi.kharkov.ua}
Abstract. The article offers a methodology for solving the problem of allo-
cating investments to optimize the work of the bank. For this, a hierarchy of cri-
teria is formed based on the use of expert information. After that, a formalized
presentation of the problem is given: how to allocate the amount of investment
according to the criteria in the optimal way. Due to the fact that the evaluation
criteria are contradictory, a utility function is built on the basis of the mathemat-
ical apparatus of fuzzy sets to solve the problem. The result of the work is a ma-
thematical model for solving the problem of distribution of investments.
Keywords: Analytic Hierarchy Process, Soft Computing, Bank Evaluation
1 The roots of the problems of bank attractiveness evaluation
Over the past few years the state of Ukraine's financial system has changed signifi-
cantly. This seriously affected banking sector. As follows from the statistics of the
Ministry of Finance [1], over the last 2 years the number of commercial banks in
Ukraine has decreased by 26%. Also, the National Bank of Ukraine implements a
policy aimed at improving the quality of banking services by raising the standards of
information security. This affects the work of banks. Therefore, today, bank execu-
tives are interested in optimizing the activities of their organizations. By optimization,
in the context of this work, we should understand the attraction of new customers and
the retention of old ones. To do this, it is necessary to understand what customers are
guided by when choosing a bank, what criteria are used, and how to invest in these
criteria with maximum efficiency.
1.1 Related works, research and publications
The problem of assessing the attractiveness of the bank can be considered from dif-
ferent points of view. In particular, it can be considered as an attraction with regard to
a consumer of banking services (order) and investment attractiveness, that is, in terms
8
of another financial institution. For instance, N. Jaremenko [2] states that an invest-
ment attractiveness of the bank is evaluated for merger or acquisition. The study pro-
poses the sequence of formed steps to evaluate the attractiveness of the bank and a
quantitative method of evaluation. For this purpose, the authors define a number of
criteria for evaluating the investment attractiveness of the bank. It should also be
noted that in some cases, the evaluation of the bank divisions (offices) is the main
subject of the work. This research evaluates the attractiveness of the branch from the
bank point of view, that is, evaluating the efficiency of the banking department. In
this regard the authors also propose a number of steps for quantitative evaluation and
suggest a group of criteria for evaluating the effectiveness of the banking department
which were developed by them.
At the same time the attractiveness of the bank from the perspective of private clients
and the attractiveness from the point of view of corporate clients are different con-
cepts because they use different criteria for the evaluation of the bank. Also objectives
of the bank attractiveness evaluation should be distinguished. Above all it can be an
attraction of new customers or retention of existing ones. The work [3] clearly states
that retaining existing customers unambiguously costs the bank cheaper and more
profitable than attracting new ones. In this work not so much the attractiveness of the
bank as the client's satisfaction with banking services is estimated. At the same time
in a number of Russian-language research (i.e., closer to the realities of Ukraine) it is
argued that the banking market is now saturated and the main struggle is being fought
for keeping customers. All the studies are united by the fact that to evaluate the attrac-
tiveness of banks the evaluation criteria are used. For instance, in [3] the RATER
model is used for service quality evaluation, which contains such groups of criteria:
─ reliability,
─ confidence (quality of service based on the ability to inspire confidence ),
─ tangibles (criteria which physically present the service),
─ emotional impression,
─ responsiveness
And also negative criteria are introduced (which reduce the level of satisfaction with
banking services).
In a number of studies, the notion of bank attractiveness is being investigated among a
specific group of clients (social, age, etc.). For example, the research [4] provides the
criteria for selecting a bank among transport workers in one of the cities of Nigeria. In
this work based on a survey of respondents, the criteria for bank evaluation were de-
fined, and their importance was also determined. The most important criteria in the
study are reliability, reputation of the bank, courtesy of the staff. And the least impor-
tant criteria for this group of customers were free cash delivery to the house, the
availability of Internet banking and the presence of an overdraft. The main criteria for
bank evaluation for transport workers will be significantly different from the criteria
of businessmen, workers, students. Research [5] contains a list of possible positions
on which a set of bank customers can be divided: this is the age segment, geographic,
behavioral, social, and so on). This survey makes the evaluation of the bank among
students of South Africa. As a result of the study it was concluded that the most im-
9
portant factors for bank evaluation for students are the ease of a bank account open-
ing, the financial stability of the bank and the location of ATMs. At the same time the
least important criteria are the availability of parking spaces, the teachers’ influence
and free gifts for customers.
It is possible to note the similarity in the studies: they all use some set of evaluation
criteria, and a mathematical model for determining the quantitative measure of the
attractiveness of the bank and determining the value criteria impact on the evaluation.
Some surveys suggest a study of the attractiveness of the bank from the customers’
point of view which is the purpose of this work. Also, several studies analyze the
attractiveness of the bank within a particular group of respondents (students, transport
workers ...). Based on the literature review and the experts’ survey, a number of crite-
ria for evaluating the bank by private clients will be selected regardless of the user
segment.
1.2 Research aims
For the first time in the work [6] an effort was made to formalize the term of the
bank’s attractiveness from the customer’s side. A model is developed in that work. On
the basis of this model an algorithm for optimizing the distribution of financial re-
sources aimed at increasing the attractiveness of the bank is proposed, since the bank
attractiveness influences in growth of its revenues. The essence of the model and the
algorithm is as follows. Based on the analytic hierarchy process, and on open statistics
and expert opinions, a hierarchy of criteria has been developed, in which the focus of
the problem is on the first level, which is formulated as a bank’s attractiveness. At the
second level there are three complex criteria on which the focus of the problem de-
pends: reliability, quality of service, range of services. Further on, the third level re-
veals the individual components of these complex indicators. For example, for the
bank's reliability criteria such indicators are: the volume of the authorized capital, the
volume of assets, the rates for deposits, and so on. At the next level of the hierarchy
there are separate alternatives. These are competing banks for which, based on the
experts' opinion, the procedure of the AHP is conducted, and their weighting factors
are determined from the point of view of the main problem focus - the attractiveness
of the bank. Next, a parametric analysis of the degree of influence of each third-level
criteria on the problem focus is made by investing in certain areas of the bank's activi-
ties: increase of the authorized capital, change rates on loans (deposits), and so on. On
the basis of this analysis, certain utility functions (bank attractiveness) are synthesized
depending on the financial resources being invested and a mathematical model of
their optimal distribution is formulated. The objective function of the model is addi-
tive. Each component determines the utility in terms of criteria for the third level of
the hierarchy. However, in [6] the concept of attractiveness does not take into account
many aspects that allow us to consider the proposed model and algorithm as some
guide to action. Therefore, the goal of this paper is to evaluate a number of factors
that will allow us to develop a method based on the model and algorithm that are
more applicable in the real situation. Let’s emphasize the main tasks that will make
possible to achieve the goal:
10
Accounting for the decision-making distribution ;
Use of fuzzy sets in Saaty's paired comparison method on which the AHP is based.
2 Mathematical model for evaluation of the bank attractiveness
2.1 Decision making distribution
Assume that the proposed method is intended for a fairly large bank which functions
at least at the state level. Then, the concept of distribution in this paper is treated as
follows. The fact is that the influence of the criteria of the third level of the hierarchy
in the AHP on the concept of a bank attractiveness is different. We introduce the con-
cept of their influence at the global, regional levels and at the level of individual clus-
ters of entities that are potential customers of a bank. In accordance with this, all the
criteria of the third level of the hierarchy are divided into four groups. The criteria
affecting the bank attractiveness:
All clusters within the entire territory of the bank's operation;
A separate type of cluster within the entire territory of a bank's operation;
All clusters within the region;
A separate cluster type within the region.
For the criteria of the first group, as previously, 4 levels of the hierarchy are used [7].
For the second and third groups, the criteria of the third level of the hierarchy give
rise to the fourth. For each criterion of the second group, these are sub-criteria consi-
dered at the level of an individual cluster type, and for the third group, these are sub-
criteria considered at the level of a particular region. Criteria of the fourth group gen-
erate the fourth and fifth levels of the hierarchy. The fourth is the level of regions, and
the fifth is the level of individual clusters of the region.
Figure 1 presents a visual interpretation of the distributed hierarchical decision sup-
k k k k
port system, where G G1 , G2 , G3 , G4 , k 1,3 is a set of criteria groups for
k
the k -th complex criterion of the third level of the hierarchy; I I1 ..I M - a set of
regions; T T1 ..TN - a set of clusters; T ( I l ), l 1, M - a set of clusters in the l -
th region, and the universal set of banks B B1 , B2 ,..., BL .
11
Fig. 1. Distribution of the decision making
2.2 Fuzzy sets in AHP
In this paper unlike the traditional approach [8] it is suggested within the AHP me-
thod to consider criteria as fuzzy sets that are defined on universal sets of variants by
means of membership function. Let's examine this approach as the fragment of the
2
example shown in Figure 1. Let the third group of criteria G3 be considered in the
complex criteria "Quality of Service", and there are such sub-criteria as "Number of
ATMs" in it. Then the criteria of the fifth level I1 , I 2 ..., I M will characterize the at-
tractiveness of the bank in the i -th region, i 1..M , and be fuzzy sets on the uni-
versal set of banks B B1 , B2 ,..., BL .
If we assume that the i ( Bl ) [0,1] characterizes the evaluation of the Bl -th bank
according to the I i -th criteria, then the fuzzy criteria
I i {[B1 , i ( B1 )],[ B2 , i ( B2 )]...[BL , i ( BL )]} , where i ( Bi ) is the member-
ship function to the fuzzy set I i . Using the Saaty’s method of paired comparisons
with the nine-point scale, i ( Bl ) numbers that characterize the rating of the l -th
*
L
bank on the i -th criteria are defined. And ( B ) 1, i 1, M . According to
l 1
*
i l
the Bellman-Zade approach [9], a fuzzy solution of the problem of achieving fuzzy
M
~
goals is determined by their intersection. That is, on a fuzzy set I I , the mem-
i
i 1
bership function of which is determined as a minimum of the membership function to
12
an individual criterion. However, the criteria I i , i 1, M have the certain impor-
M
tance. Suppose that it is determined by the weighting coefficients i 0, 1
i 1
i
, which values can be determined by the method of paired comparisons using the Saa-
ty nine-point scale. Then the fuzzy set of intersection of fuzzy goals has the formula
(1) and the alternative with the maximum value of the membership function is consi-
dered to be the best.
~
I {[B1 , min( *i ( B1 )) i ], [ B 2 , min( *i ( B2 )) i ],...,[ B L , min( *i ( BL )) i ]} (1)
3 Method for increasing bank attractiveness in terms of limited
resources
An offered method is based on the results of research [5] and on this survey. It is pre-
sented as a sequence of steps.
First step. A set of lower-level criteria X is formed based on the hierarchy (Fig.1).
This set characterizes bank attractiveness from the customer’s point of view. Each
criterion has certain indicators which affect bank attractiveness. For instance, in group
3 (Range of Services) there is a criterion “Remote banking services”. It has such
indicators as availability of mobile applications (for different OS: Android, iOS, WP,
etc.), availability of remote banking service for companies, quality of these services,
level of support responsiveness, etc.
Second step. For each criterion from X , an evaluation of indicators which character-
ize an attractiveness of the bank from B (a set of alternatives – different banks) is
made by experts. Assume that for each l -th bank, l 1, L , a set of these evaluations
*
is X l .
Third step. Based on sets X l , l 1, L , and hierarchy of criteria used AHP (Section
*
2.2) importance of weighting factors l ( X l ) are formed. These factors are formed
*
from their attractiveness point of view, and satisfy l ( X l ) 0, l 1, L. Also
*
L
( X ) 1.
l1
l
*
l
*
Fourth step. Variation of evaluations X l from the set X to increase the l -th bank
importance of weighting factor requires certain financial costs. Let y l j financial costs
of l -th bank used for the correction of j -th evaluation xl*j X l* , and ml j - limit
value of y l j which is determined by the bank experts.
13
Fifth step. For j -th criteria of l -th bank a set of experiments is made. In these expe-
riments y l j varies in the range 0 yl j ml j , and the importance weighting factor
* * *
l ( X l ) based on AHP is determined, where X l {xl*1 ,..., x l j ,..., xl*S } , S X ,
*
and x l j is a corrected evaluation of indicator of j -th criteria of l -th bank connected
with financial costs y l j .
Sixth step. A function of the impact of financial costs for the correction of indicators
j -th criteria of l -th bank on the importance of weighting factor of l -th bank is
formed based on a set of experiments.
*
l ( X l ) Fl ( yl ) j j
(2)
Seventh step. After the analysis of function Fl j ( yl j ) , some subsets of criteria are
ignored, which are not sufficient or do not have any influence on the importance of
weighting factor (2). Next step will deal with a subset X X of existing criteria.
C
Eighth step. Let K is an amount of costs which the bank can spend for increasing its
attractiveness from the customer’s point of view. Then the rational distribution of
these costs is determined on the basis of the solution of the following problem: find
such value of vector yl { yl j } , which ensures maximum value of target function:
*
Fl ( yl ) Fl j ( yl j ) (3)
jX C
, with conditions:
0 yl j ml j , j X C (4)
y K. lj (5)
jX C
Based on solution of (3-5), the rational distribution of financial costs K is determined.
This distribution ensures the increasing of a bank attractiveness from the customers’
side.
In this research we assumed that the function (3) is an additive function to functions
(2).
14
4 Conclusions and future work
The paper describes the process of evaluation and a bank development management
in terms of its attractiveness using the distributions of this concept at the level of indi-
vidual regions and user clusters, and offers method for amount distribution for in-
creasing attractiveness based on AHP. The problem of the application of soft compu-
tations in the AHP is considered due to the "blurring" of the criteria for all levels of
the hierarchy and the introduction of fuzzy sets at different levels. This approach is an
alternative to the traditional AHP algorithm. Further research will be devoted to com-
paring the traditional AHP algorithm, and based on soft computing, as well as the
validity of applying one or another approach in different situations.
References
1. Ministry of Finance of Ukraine, last banking statistics,
https://index.minfin.com.ua/bank/stat/count.php, last accessed 2018/01/08.
2. Yaremenko N.: Bank investment attractiveness and the methodology for its assessment at
mergers and acquisitions. MPRA Paper №60635, pp.412-420 (2014).
3. Titko E., Lace N.: Customer satisfaction and loyalty in Latvian retail banking. Economics
and management journal, 2010.15, pp.1031-1038 (2010).
4. Nwaizugbo I.: Factors influencing the choice of commercial bank among transport work-
ers in Anambra state Nigeria. British Journal of Marketing Studies, 4(6), pp.12-25, (2016).
5. Chigamba C., Fatoki O. Factors influencing the choice of commercial banks by university
students in South Africa. International Journal of Business and Management 6(6), pp.66-
76 (2011).
6. Godlevskyi, M., Dabahian D.: Model for optimizing the attractiveness of a bank in condi-
tions of limited resources. Bulletin of NTU “KhPI”. Series: System analysis, control and
information technology 28(1250), pp.79–83 (2017).
7. Saaty T.L.: The Analytic Hierarchy Process. McGraw Hill, International, New York
(1980).
8. Abdikeeva N.: Cognitive Business Analytics. INFRA-M, Moscow (2011).
9. Raskin L., Seraya O.: Fuzzy mathematics. Foundations of the theory. Annexes. Parus,
Kharkov (2008).
15