See discussions, stats, and author profiles for this publication at: [Link]

net/publication/363484486

A Novel TOPSIS Method under Fermatean Fuzzy Hypersoft Set Based on

Correlation Coefficients for Selection of Hip Prothesis Materials

Preprint · September 2022

CITATIONS READS

0 35

1 author:

Murat Kirişci
İstanbul University-Cerrahpaşa
128 PUBLICATIONS   667 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

[Link] View project

[Link] View project

All content following this page was uploaded by Murat Kirişci on 12 September 2022.

The user has requested enhancement of the downloaded file.

A Novel TOPSIS Method under Fermatean Fuzzy Hypersoft Set Based
on Correlation Coefficients for Selection of Hip Prothesis Materials
Murat Kirişci

Istanbul University-Cerrahpaşa, Cerrahpasa Medicine Faculty,

Department of Biostatistics and Medical Informatics, Istanbul, Turkey
e-mail: [Link]@[Link]

Abstract: A statistical technique called correlation analysis is used to determine the

link between two variables and gauge how strongly two variables are linearly related. The
degree of change in one variable as a result of the other’s change is determined via correlation
analysis. The correlation coefficient is one of the statistical ideas that are most relevant to
this kind of investigation. The correlation coefficient, which is easily recognized by its sym-
bol r and typical value without units falling between 1 and -1, is the unit of measurement
used to determine the intensity of the linear relationship between the variables included in
correlation analysis. One of the best at representing completely ambiguous and uncertain
information is the Fermatean fuzzy set. In this research, we present correlation coefficients
and weighted correlation coefficient formulations to evaluate the relationship between two
Fermatean fuzzy hypersoft sets, taking into account that the correlation coefficient plays
a significant role in statistics and engineering disciplines. The Fermatean fuzzy hypersoft
set is a parameterized family that deals with the subattributes of the parameters and is
an appropriate extension of the Fermatean fuzzy soft set. It is also the generalization of
the intuitionistic fuzzy hypersoft set and the Pythagorean fuzzy hypersoft set, which are
used to accurately assess insufficiency, anxiety, and uncertainties in decision-making. The
Fermatean fuzzy hypersoft set can accommodate more uncertainties compared to the in-
tuitionistic fuzzy hypersoft set and Pythagorean fuzzy hypersoft set, and it is the most
substantial methodology to describe fuzzy information in the decision-making process. One
of the aims of this study is to give Fermatean fuzzy hypersoft sets and to examine their
basic properties. The second objective of this study is to develop the notion and features of
the correlation coefficient and the weighted correlation coefficient for the Fermatean fuzzy
hypersoft set and to introduce the aggregation operators such as Fermatean fuzzy hyper-
soft weighted average and Fermatean fuzzy hypersoft weighted geometric operators under
the Fermatean fuzzy hypersoft set scenario. A prioritization technique for order preference
by similarity to the ideal solution (TOPSIS) under Fermatean fuzzy hypersoft set based
on correlation coefficients and weighted correlation coefficients is presented. Through the
developed methodology, a technique for solving the multi-attribute group decision-making
problem is planned. In addition, examples of medical decision-making are presented for the
importance and application of the developed methodology.

Keywords: Fermatean fuzzy hypersoft set, correlation coefficients, informational en-

ergy, group decision-making, TOPSIS, medical decision-making.

1 Introduction
1.1 Multi-criteria decision-making and uncertainty
Cognitive science, philosophy, artificial intelligence, and psychology are just a few of the
academic disciplines that study how people think and make decisions in everyday situa-
2

tions. Several mathematical and statistical models are frequently used to characterize these
processes. The issue of decision-making (DM) emerges during this procedure. DM is the
process of choosing one or more of the available possibilities for action when a person or
organization is striving to achieve a specified objective. In real life, although many every-
day decisions can be made instinctively, it is a well-known fact that complex and critical
decisions require more thought. For making such complex decisions, Multi-Criteria Deci-
sion Making (MCDM) is a set of analytical tools that evaluate the benefits and drawbacks
of various options based on various criteria. MCDM techniques are used to select one or
more alternatives from a collection of alternatives with varying characteristics according to
competing criteria or to rank these alternatives to aid the DM process. In other words,
decision-makers use MCDM techniques to rank options based on various characteristics by
comparing them to multiple criteria. In other words, MCDM is a set of procedures that are
used every day at all levels and in all aspects of life.

MCDM is based on the handle of modeling the decision procedure using criteria and ana-
lyzing the decision maker in such a way that the benefit obtained at the end of the process is
maximized. Assume you are in the process of making a home purchase decision. You should
be asked to evaluate the two house options available to you based on two criteria, such as
price and house size, and make a decision. In such a decision problem, a simple comparison
can be used to select the alternative that will provide you with the greatest benefit (the
house with the greatest area at the lowest price) without resorting to any decision-making
method. However, if we apply the problem of buying a house to real life, we can predict
that the number of alternatives will be much greater than two, and the number of criteria
you must consider during the buying process will be much greater. In this case, you will be
unable to conduct logical tests in order to make an effective decision, and making decisions
based on intuition will not yield very productive results. In such a case, utilizing a scientific
decision support system to aid your decision-making process will allow you to obtain an
effective output as a result of the DM process. The decision problem described above on a
simple example becomes more complex in real life, particularly in the decision processes of
businesses, hospitals, and disease diagnosis, economies, and government plans, and becomes
more important in terms of costs when it is resolved. The MCDM approaches proposed for
use in this type of decision problem include approaches and methods that attempt to find the
”best/suitable” solution that meets more than one conflicting criterion. The best solution
is the decision made as a result of the decision process that provides the greatest benefit
at the lowest cost. To overcome such issues, decision-makers can use MCDM techniques to
make deterministic and, thus, more effective decisions. MCDM is a collection of strategies
that are widely employed in all aspects of life and at all levels.

Uncertainty is a key notion in DM problems. Unpredictable events characterize uncer-

tainty. In uncertain circumstances, routine decisions cannot be debated. In ambiguous
situations, it is critical to weigh both the benefits and drawbacks of potential outcomes.
It is critical at this point to conduct a thorough examination of the environmental factors.
Even when future judgments are uncontested, drawing on prior experiences and decisions is
not always successful when there is uncertainty. Because of Zadeh’s concept of ”fuzzy sets”
(FS) [51], everyday linguistic phrases have become ”computable.” Thanks to fuzzy logic, the
grading system was able to expand the area of classical mathematics that was previously
limited to certainty. As a result of its effective implementation in everyday situations, this
concept prompted a paradigm change that expanded throughout the world. A distinct func-
tion element is either an element of a set in the conventional sense or it is not. According
to the FS concept, the membership function (MF), which assigns each item a degree of
membership in the range [0, 1], determines whether or not an apple is a member of a set.
 3

The degree to which an element belongs to a set in FS A is ρ(A), while the degree to
which it does not belong is 1−ρ(A). However, in a number of cases, this condition falls short
of fully addressing the uncertainty. As a result, Atanassov [1] developed the intuitionistic
fuzzy set (IFS) theory as an extension to the FS theory. In addition to the membership
degree (MD), IFS theory specifies the non-membership degree (ND), whereas FS theory
only reveals the membership degree (MD)(∈ [0, 1]). In the theory of IFS, MD and ND are
both in the [0, 1] range. Yager [46] proposed Pythagorean fuzzy sets (PFS) as an extension
of IFSs in some cases because IFSs cannot adequately represent uncertainty. PFSs are based
on the idea that M D2 +N D2 ≤ 1. There is a lot of study on FS and its numerous expansions
in the literature([6, 7, 8, 9, 11, 13, 14, 32, 33, 50]).

1.2 Correlation Coefficients

Any statistical relationship, whether causal or not, between two random variables or bivari-
ate data is referred to as correlation or dependence. Although the term ”correlation” can
apply to any sort of link, it is most commonly used in statistics to describe the degree to
which two variables are linearly connected. As a result, in correlation research, the correla-
tion coefficient (KK) is the exact measurement used to determine the power of the linear link
between 2 variables. Correlations are useful because they can forecast a relationship that can
be used in practice. In general, the appearance of a correlation does not imply the presence
of a causal relationship (i.e., a correlation does not imply causation). If random variables
do not meet a mathematical characteristic known as probabilistic independence, they are
called dependent. Correlation is often used interchangeably with dependence. When used
in a technical sense, correlation refers to any of many distinct types of mathematical op-
erations conducted between the tested variables and their corresponding predicted values.
A correlation is a measure of co-change. We can use this statistical method to determine
the strength and direction of a relationship between two or more variables. The letter r
represents the letter KK. According to mathematics, the KK can have values ranging from
−1 to +1. The direction of the link between the two variables is indicated by the sign of
this coefficient and the degree of correlation between the two variables. Consider the recip-
rocal connection to gain an understanding of the correlation. A correlation is a reciprocal
relationship that can represent the difference or resemblance between two variables rather
than a cause-and-effect relationship. Because of correlation, if the relationship between the
variables is known, the other can be estimated by studying the value of one variable, or the
values of the connected variables can be maximized if the influencing factors are managed.

When there are more than two variables, a KK is a multivariate statistic; when there
are only two variables, it is a bivariate statistic. As a result, several fields of study have
emerged, ranging from engineering to physics, and from medicine to economics. Existing
probabilistic approaches have several benefits but also drawbacks. Probabilistic techniques,
for example, achieve the required confidence level by collecting large amounts of random
data. The complex system, on the other hand, has widespread fuzzy uncertainty, making it
impossible to predict every possible outcome of the events. Because specialists can only act
on numerical information, probabilistic outcomes may not always give helpful information.
Furthermore, there are times when there is insufficient data to operate parameter statistics
appropriately in day-to-day operations. Because of these constraints, probability theory’s
conclusions may not always ensure beneficial information to specialists, and the probabilistic
approximation is thus insufficient to account for the underlying uncertainties in the data.
There are several approaches to overcoming these obstacles. Methods based on FS theory
are among the most successful outcomes of these alternatives for reducing uncertainties and
4

imprecision in DM.

Finding a KK between any two parameters or variables is a fairly common statistical

task. Pearson’s KK has been used in data processing and categorization, pattern recog-
nition, clustering, medical diagnosis, and decision-making statistics research. It has been
demonstrated that classical correlation cannot handle ambiguous data. To address and
quantify ambiguities in human perception and reasoning, the fuzzy logic approach is used.
Scientists are accustomed to analyzing data using binary logic. Because human logic is
imprecise and complicated, employing binary logic to study human cognitive processes re-
sults in some distortion. The foundation of fuzzy logic is human cognitive processes. The
modeling of thinking and decision mechanisms that enables people to reach consistent and
correct conclusions in the face of incomplete and inaccurate information is one definition
of fuzzy logic. The fuzzy type KKs have been extended with the help of mathematical
statistics, statistical KKs, and the fuzzy logic method. The KK generated for fuzzy data
not only displays the degree of association between FSs, but also whether they are favorably
or negatively connected.

1.3 Motivation
In decision-making situations, tools such as aggregation operators and information mea-
sures are routinely employed. The KKs assessment of the amount of dependence between
two sets may also be used to choose the optimal alternative. Using KKs, one may deter-
mine how strongly two variables are related. Because the information in various settings
is typically unclear, ambiguous, and partial, numerous scholars have created KKs in fuzzy
environments. Chiang and Lin [2] developed a strategy for KK of FSs in addition to pro-
viding the correlation for fuzzy information in line with traditional statistics. The KK of
fuzzy information has been examined in [24] using a mathematical programming approx-
imation, according to the standard notion of KKs. According to the findings of the FS
theory, Atanassov [1] claimed that IFS results were more complete and precise. The IFS
theory considers both MD and ND, and it demands that their sum be one or less than one.
The IFS-derived KKs have a variety of applications, including DM, cluster analysis, image
processing, pattern recognition, and so on ([23], [41], [42], [43], [44]). Several DM problems
utilizing Pythagorean fuzzy information have been described in the literature as a result of
the PFS ([13], [12], [31], [32], [45], [46], [47], [53]), which was designed to address an IFS issue.

Many area experts and scholars have expressed interest in the PFS model since its
creation. For example, Yager and Abbasov, for example, examined the relationship be-
tween Pythagorean MGs and complex numbers. Zhang and Xu developed an extended
Pythagorean fuzzy TOPSIS model to address various MCDM situations utilizing Pythagorean
fuzzy data. Furthermore, it offers a wide range of possible applications, including domestic
airline service quality, DM, and so on. However, in other cases, the PFS strategy could not
be approved. Consider an expert team that was separated into two groups. The first team
of experts estimates the MD to be 0.9, whereas the second team estimates the ND to be 0.8.
It is obvious that 0.92 + 0.82 is larger than one. The IFS and PFS were unable to depict
this circumstance.

Senapati and Yager [34] developed the concept of FFSs as an extension of the IFSs and
PFSs to address this difficulty. In an FFS, the cubic sum of an object’s MD and ND is
limited by 1 (0 ≤ M D3 + N D3 ≤ 1). Consider 0.9 + 0.6 > 1, (0.9)2 + (0.6)2 > 1, and
(0.9)3 + (0.6)3 < 1 as examples. That is because the total of the cubes of the MD and
ND of FFSs is in the [0, 1], FFSs give a more comprehensive view for FSs. When dealing
 5

with unclear data, FFSs are more adaptable and efficient than IFSs and PFSs. FFS is now
playing an important role in a variety of disciplines since it is a strong notion for dealing
with imprecise and unclear information in a Fermatean fuzzy environment.

Some FFS characteristics, score, and accuracy functions are provided in [34]. Addition-
ally, the TOPSIS approach, which is widely used to solve MCDM issues, has been employed
to solve FFS. Additionally, Senapati and Yager [34] have used the TOPSIS method, which
is often used in MCDM issues, to solve FFS difficulties. Senapati and Yager [35] continued
this work by investigating a number of additional operations including arithmetic mean op-
erations over FFSs in addition to using the FF weighted product model to address MCDM
issues. New aggregation operations that are FFS-related have been described and their as-
sociated attributes have been looked at in [36].

Shahzadi and Akram [37] developed the new aggregated operators and a decision support
algorithm for the FFSS. Garg et al. [10] introduced novel FFS types of aggregated opera-
tors defined by t-norm and t-conorm. In their investigation, Donghai et al. [3] propose the
notion of FF linguistic word sets. These sets’ operations, scoring, and accuracy functions
were provided. In [4], a new similarity metric for FF linguistic word sets is developed. The
new metric is a hybrid of the Euclidean distance measure and the cosine similarity measure.
Kirisci [20] designed FF soft sets and provided an entropy metric based on them. In [18],
a novel hesitant fuzzy set known as the ”fermatean hesitant fuzzy set” is presented and
its properties are studied. Kirisci and Simsek [19] offer aggregation operations to extend
FFHSs to interval-valued Fermatean hesitant fuzzy sets (IVFHFS) and to improve MCGDM
methods to IVFHF environments. The ELECTRE I approach is defined in [15] using Fer-
matean fuzzy sets and the group DM process, in which more than one individual engages at
the same time. [17] defines many FF reference relations (consistent, incomplete, consistent
incomplete, acceptable incomplete). A priority vector-based additive consistency has been
provided. In addition, a methodology for obtaining missing judgments in incomplete FF
preference relations is described. Garg et al. [10] examined decision-making analysis using
Fermatean fuzzy Yager aggregation operators (for use in coronavirus disease 2019 (COVID-
19) testing facilities). Yang et al. [49] addressed Fermatean fuzzy function differential cal-
culus. Shahzadi and Akram [37] created a unique DM paradigm based on Fermatean fuzzy
soft information to pick an antiviral mask. Furthermore, in a Fermatean fuzzy environment,
Akram et al. [38] suggested an unique DM approach for selecting an efficient sanitizer to
decrease COVID-19. Based on the linguistic scale function, Liu et al. [25] established the
idea of distance measures for Fermatean fuzzy linguistic word sets. Their application in the
creation of TODIM and TOPSIS approaches exemplifies this.

The originality:
Fuzzy KK, IF KK, and PF KK are examples of expansions to the classical KKs. The
KKs’ performance has increased as a result of these extensions. FFSs can handle ambigu-
ity and partial information problems more effectively than IFSs and PFSs (Figure 1). In
this study, first of all, the concept of the FFHS is introduced and its basic properties are
examined. Secondly, new KKs based on FFHSS were defined and the theoretical basis of
these coefficients was demonstrated. The reason for using FFSs to define new correlation
coefficients is that since the M D3 + N D3 ≤ 1 requirement for an object is met, it is likely to
cover more items than IFSs and PFSs. The MCGDM algorithm was produced by combining
the new KKs and the TOPSIS method, and an example for the selection of hip prosthesis
materials is given to demonstrate the operability and reliability of the method. KKs given
in previous studies and KKs defined in this study were compared.
6

Figure 1: Comparison of FFSs, PFSs and IFSs [34]

The organization of this work’s structure: Fundamental ideas and findings from
the concept of HSSs are given in Section 2. The new FFHSSs, which are an extension of the
FFSS and also a parameterized family that deals with the sub-attributes of the parameters
are given in the third chapter, and their properties are examined in detail. The informational
energies, correlations, KKs, and weighted KKs of FFHSSs are defined in Section 4. Section
5 presents a prioritizing strategy based on KKs and weighted KKs for order preference by
the resemblance to the ideal solution (TOPSIS) under FFHSS. An approach for tackling
MCGDM issues is envisaged using the described approach. The 6th chapter is based on
an illustrative example of our new method. In this section, a study was carried out on
the selection of hip prosthesis materials. A comparison of the new method with previous
methods is made in Section 7 and this section is concluded with a subsection explaining the
superiority of the new method.

2 Preliminaries
Throughout the article, U, and E will be used as a universe and as a set of attributes,
respectively.
Definition 2.1. [28] FA is called a SS over U, where F : A → P (U), A ⊆ E, and P (U)
is the power set of U.
SS can also be written as:
FA / A }.
= {F (e) ∈ P (U) : e ∈ ϵ, F (e) = ∅, if e ∈ (1)
Maji et al. [26] investigated SS and FS and proposed a more extended form to handle
uncertainty in comparison to current FS and SS, as well as its unique properties. This is
sometimes referred to as a fuzzy soft set (FSS), which is a mix of FS and SS.
Definition 2.2. [26] FA is called a FSS over U, where F : A → F (U), A ⊆ E, and
P (U) is a collection of all fuzzy subsets of U.
Definition 2.3. [40] Let α = {α1 , α2 , · · · , αn } (n ≥ 1) be a set of attributes, and set
Σi be a set of corresponding sub-attributes of αi , respectively, with Σi ∩ Σj = φ for n ≥ 1
∀i, j ∈ {1, 2, · · · , n}, and i ̸= j. Consider Σ1 ×Σ2 ×· · ·×Σn = A = {m1p ×m2r ×· · ·×mnx }
be a collection of multi-attributes, (1 ≤ p ≤ a, 1 ≤ r ≤ b, and 1 ≤ s ≤ c and a, b, c ∈ N),
and P (U) is a power set of U. Therefore, FΣ1 ×Σ2 ×···×Σn =A is called a HSS, where
F : Σ1 × Σ2 × · · · × Σn = A → P (U). (2)
HSS can also be written as:
(m̃, F (A )(m̃)) : m̃ ∈ A , F (A )(m̃) ∈ P (U) .

FA = (3)
 7

Definition 2.4. Let α = {α1 , α2 , · · · , αn } (n ≥ 1) be a set of attributes, and set Σi

be a set of corresponding sub-attributes of αi , respectively, with Σi ∩ Σj = φ for n ≥ 1
∀i, j ∈ {1, 2, · · · , n}, and i ̸= j. Consider Σ1 ×Σ2 ×· · ·×Σn = A = {m1p ×m2r ×· · ·×mnx }
be a collection of multi-attributes, (1 ≤ p ≤ a, 1 ≤ r ≤ b, and 1 ≤ s ≤ c and a, b, c ∈ N),
and P U is a collection of all fuzzy subset of U. Therefore, FΣ1 ×Σ2 ×···×Σn =A is said to be
fuzzy HSS(FHSS), where

F : Σ1 × Σ2 × · · · × Σn = A → P U . (4)

FHSS can also be written as:

(m̃, F (A )(m̃)) : m̃ ∈ A , F (A )(m̃) ∈ P U .


FA = (5)

Definition 2.5. Let α = {α1 , α2 , · · · , αn } (n ≥ 1) be a set of attributes, and set Σi

be a set of corresponding sub-attributes of αi , respectively, with Σi ∩ Σj = φ for n ≥ 1
∀i, j ∈ {1, 2, · · · , n}, and i ̸= j. Consider Σ1 ×Σ2 ×· · ·×Σn = A = {m1p ×m2r ×· · ·×mnx }
be a collection of multi-attributes, (1 ≤ p ≤ a, 1 ≤ r ≤ b, and 1 ≤ s ≤ c and a, b, c ∈ N),
and IF S U is a collection of all intuitionistic fuzzy subset of U. Therefore, FΣ1 ×Σ2 ×···×Σn =A
is said to be intuitionistic fuzzy HSS(IFHSS), where

F : Σ1 × Σ2 × · · · × Σn = A → IF S U . (6)

It can be also given as

(m̃, F (A )(m̃)) : m̃ ∈ A , F (A )(m̃) ∈ IF S U .


FA = (7)

where F (A )(m̃)) = λ, ζF (m̃) (λ), ηF (m̃) (λ) : δ ∈ U , in which ζF (m̃) (λ) and ηF (m̃) (λ) show


the MV and NV of the attributes such as ζF (m̃) (λ), ηF (m̃) (λ) ∈ [0, 1], and 0 ≤ ζF (m̃) (λ) +
ηF (m̃) (λ) ≤ 1.

For L = {(x, y) : x, y ∈ [0, 1], x3 + y 3 < 1}, let (L , ⪯L ) denotes a complete lattice,
where ⪯L is the corresponding partial order with (x, y) ⪯L (m, n) ⇔ x ⪯ mandy ⪯ n
for all (x, y), (m, n) ∈ L . The ordered pair (x, y) ∈ L is said to be FF value(FFV) or FF
number(PFN).

According to the definition of complete lattice, the FFS can be given as:

Let D : U → L be a L -fuzzy set. Then, the FFS D(a) = {(a, ζD (a), ηD (a)) : a ∈ U}
can be identified as D(a) = (ζD (a), ηD (a)) for all a ∈ U.

3 Fermatean Fuzzy Hypersoft Sets

In this section, we present FFHSS and some fundamental operations with their characteris-
tics under the FFHS environment.

Definition 3.1. Let α = {α1 , α2 , · · · , αn } (n ≥ 1) be a set of attributes, and set Σi

be a set of corresponding sub-attributes of αi , respectively, with Σi ∩ Σj = φ for n ≥ 1
∀i, j ∈ {1, 2, · · · , n}, and i ̸= j. Consider Σ1 ×Σ2 ×· · ·×Σn = A = {m1p ×m2r ×· · ·×mnx }
be a collection of multi-attributes, (1 ≤ p ≤ a, 1 ≤ r ≤ b, and 1 ≤ s ≤ c and a, b, c ∈ N),
8

and F F S U is a collection of all fermatean fuzzy subset of U. Therefore, FΣ1 ×Σ2 ×···×Σn =A
is said to be Fermatean fuzzy HSS(FFHSS), where

F : Σ1 × Σ2 × · · · × Σn = A → F F S U . (8)

.
It also be defined as:

(m̃, F (A )(m̃)) : m̃ ∈ A , F (A )(m̃) ∈ F F S U .


FA = (9)

where F (A )(m̃)) = λ, ζF (m̃) (λ), ηF (m̃) (λ) : δ ∈ U , in which ζF (m̃) (λ) and ηF (m̃) (λ) show


the MV and NV of the attributes such as ζF (m̃) (λ), ηF (m̃) (λ) ∈ [0, 1], and 0 ≤ ζF (m̃) (λ)3 +
ηF (m̃) (λ)3 ≤ 1.

Example 3.2. Let U = {λ1 , λ2 } be the universe of discourse.

The collection of attributes denoted by L = {technology methodolgy, subjects, classes} =
{L1 , L2 , L3 }. Their corresponding attribute values given as

L1 = {project base, class discussion} = {m11 , m12 }

L2 = {M athematics, ComputerScience, Statistics} = {m21 , m22 , m23 }
L3 = {M asters, Doctoral} = {m11 , m12 }

Let A = L1 × L2 × L3 be a set of attributes:

A = L1 × L2 × L3 = {n11 , n12 } × {n21 , n22 , n23 } × {n11 , n12 }

(
= (n11 , n21 , n31 ), (n11 , n21 , n32 ), (n11 , n22 , n31 ), (n11 , n22 , m32 ), (m11 , m23 , m31 ), (m11 , m23 , m32 ),
)
(m12 , m21 , m31 ), (m12 , m21 , m32 ), (m12 , m22 , m31 ), (m12 , m22 , m32 ), (m12 , m23 , m31 ), (m12 , m23 , m32 )

= {m̃1 , m̃2 , m̃3 , m̃4 , m̃5 , m̃6 , m̃7 , m̃8 , m̃9 , m̃10 , m̃11 , m̃12 }.

Then, the FFHSS over U is given as follows:

n
FA = (m̃1 , (λ1 , (0.7, 0.4)), (λ2 , (0.4, 0.6))) , (m̃2 , (λ1 , (0.3, 0.8)), (λ2 , (0.2, 0.6))) ,
(m̃3 , (λ1 , (0.4, 0.8)), (λ2 , (0.4, 0.6))) , (m̃4 , (λ1 , (0.3, 0.7)), (λ2 , (0.3, 0.8))) ,
(m̃5 , (λ1 , (0.5, 0.5)), (λ2 , (0.3, 0.7))) , (m̃6 , (λ1 , (0.3, 0.8)), (λ2 , (0.2, 0.8)))
(m̃7 , (λ1 , (0.3, 0.6)), (λ2 , (0.5, 0.3))) , (m̃8 , (λ1 , (0.5, 0.7)), (λ2 , (0.3, 0.2))) ,
(m̃9 , (λ1 , (0.5, 0.7)), (λ2 , (0.4, 0.8))) , (m̃10 , (λ1 , (0.5, 0.7)), (λ2 , (0.8, 0.3))) ,
o
(m̃11 , (λ1 , (0.5, 0.7)), (λ2 , (0.4, 0.6))) , (m̃12 , (λ1 , (0.3, 0.7)), (λ2 , (0.3, 0.7)))

Remark. (1) If both (ζF (m̃) (λ))3 + (ζF (m̃) (λ))3 ≤ 1 and (ζF (m̃) (δ))2 + (ζF (m̃) (δ))2 ≤ 1
hold, then FFHSS as reduced to PFHSS [55].
(2) If (ζF (m̃) (λ))3 + (ζF (m̃) (λ))3 ≤ 1 and each parameter of the set of attributes consists
of no sub-attribute, then FFHSS was reduced to FFSS [20].
(3) If both (ζF (m̃) (λ))3 + (ζF (m̃) (λ))3 ≤ 1 and (ζF (m̃) (λ))2 + (ζF (m̃) (λ))2 ≤ 1 hold and a
set of attributes contains only one parameter with no sub-attributes, then FFHSS was
reduced PFSS [31].
 9

Definition 3.3. Let FA and GB be two FFHSS over U.

(i.) For ζF (m̃) (λ) ≤ ζG(m̃) and ηF (m̃) (λ) ≤ ηG(m̃) , λ ∈ U, if

a. A ⊆ B,

b. FA (ã)(λ) ⊆ GB (ã)(λ) for all λ ∈ U,

then FA is called a FFHS subset of GB .

(ii.) If ζF (m̃) (λ) = 0, and ηF (m̃) (λ) = 1 for all m̃ ∈ A and λ ∈ U, (∅A = {m̃, (λ, (0, 1)) :
λ ∈ U, m̃ ∈ A } is called empty FFHSS(denoted by ∅F (m̃) (λ)).

(iii.) If ζF (m̃) (λ) = 1, and ηF (m̃) (λ) = 0 for all m̃ ∈ A and λ ∈ U , (EA =
{m̃, (λ, (1, 0)) : λ ∈ U, m̃ ∈ A } is called universal FFHSS(denoted by EF (m̃) (λ)).

(iv.) If for all λ ∈ U and m̃ ∈ U, ζFA (m̃) (λ) = ζGA (m̃) (λ) and ηFA (m̃) (λ) = ηGA (m̃) (λ),
then FA and GB is called equal FFHSS.

Theorem 3.4. Let FA , GB and HC be three FFHSS over the universe U. Hence

(i.) FA ⊆ EA ,

(ii.) ∅A ⊆ FA ,

(iii.) FA ⊆ GB and GB ⊆ HC ⇒ FA ⊆ HC .

Proof. (i.) FA ⊆ EA , since ζFA (m̃) (λ) ≤ ζGA (m̃) (λ) = 1 and ηFA (m̃) (λ) ≥ ηGA (m̃) (λ) = 0
∀m̃ ∈ A , λ ∈ U.
(ii.) ∅A ⊆ FA , since 0 = ζFA (m̃) (λ) ≤ ζGA (m̃) (λ) and 1 = ηFA (m̃) (λ) ≥ ηGA (m̃) (λ) ∀m̃ ∈ A ,
λ ∈ U.
(iii.) FA ⊆ g ⇒ ζFA (m̃) (λ) ≤ ζGB (m̃) (λ) and ηGB (m̃) (λ) ≥ ηFA (m̃) (λ) ∀m̃ ∈ A , λ ∈ U.
Therefore, ζGB (m̃) (λ) ≤ ζHC (m̃) (λ) and ηHC (m̃) (λ) ≥ ηGB (m̃) (λ) ∀m̃ ∈ A, λ ∈ U. Hence, we
have FA ⊆ HC .

Definition 3.5. Let FA = {(m̃, [λ, ζF (m̃) (λ), ηF (m̃) (λ)] : λ ∈ U) : m̃ ∈ A } be a FFHSS
c
over U. FA = {(m̃, [λ, ηF (ñ) (λ), ζF (m̃) (λ)] : λ ∈ U) : m̃ ∈ A } is called complement of FA .

Proposition 3.6. If FA be a FFHSS, then

c c
i. (FA ) = FA
c
ii. ∅A = EA
c
iii. EA = ∅A

Proof. We will only prove the item (i.). The items (ii.) and (iii.) can be easily proved as
similar to (i.)

Let FA = {(m̃, [λ, σF (m̃) (λ), τF (m̃) (λ)] : λ ∈ U) : m̃ ∈ A } be a FFHSS over U. Then, the
using Definition 3.5, we have FA c
= {(m̃, [δ, τF (m̃) (λ), σF (m̃) (λ)] : λ ∈ U) : m̃ ∈ A }. Again
using the Definition 3.5, (FA ) = {(m̃, [λ, σF (m̃) (λ), τF (m̃) (λ)] : λ ∈ U) : m̃ ∈ A }. Hence
c c
c c
(FA ) = FA .
10

Definition 3.7. Let FA and GB be two FFHSS over U. The union of FA and GB is
given as:

˜ GB = HC
FA ∪ and C = A ∪ B, (10)

˜ ∈A −B

 F (m̃) , if ma
ζH(m̃) (λ) = G(m̃) , if m̃ ∈ B − A (11)
max{F (m̃), G(m̃)} , if m̃ ∈ A ∩ B


F (m̃) , if m̃ ∈ A − B


ηH(m̃) (λ) = G(m̃) , if m̃ ∈ B − A (12)
min{F (m̃), G(m̃)} , if m̃ ∈ A ∩ B


Definition 3.8. Let FA and GB be two FFHSS over U. The intersection of FA and GB
is given as:

˜ GB = HC )
FA ∩ and C = A ∩ B, (13)

F (m̃) , if m̃ ∈ A − B


ζH(m̃) (λ) = G(m̃) , if m̃ ∈ B − A (14)
min{F (m̃), G(m̃)} , if m̃ ∈ A ∩ B


F (ã) , if ã ∈ A − B


ηH(ã) (δ) = G(ã) , if ã ∈ B − A (15)
max{F (ã), G(ã)} , if ã ∈ A ∩ B


Definition 3.9. Let A , B ⊆ E, C = A ∪H ̸= ∅, F ⊆ G , and FA ⊆ GB . The restricted

˜ r GB = HC such that
union of FA and GB is characterized by (FA ∪

H = FA ∪r GB
ζH(m̃) (λ) = max{ζF (m̃) (λ), ζG(m̃) (λ)};
ηH(m̃) (λ) = min{ηF (m̃) (λ), ηG(m̃) (λ)}.

Definition 3.10. Let A , B ⊆ E, C = A ∪ H ̸= ∅, F ⊆ G , FA ⊆ GB . The restricted

˜ r GB = MD such that
union of FA and GB is defined as FA ∩

MD = FA ∩r GB
ζM (m̃) (λ) = min{ζF (m̃) (λ), ζG(m̃) (λ)};
ηM (m̃) (λ) = max{ηF (m̃) (λ), ηG(m̃) (λ)}.

Definition 3.11. For two FFHSS FA and GB , the ”OR” operation on FA and GB is
described as

FA ∨ GB ) = HoverlineA ×B = (K, ã × b̃)

where

(K , m̃ × ñ) = FA (m̃) ∪ GB(ñ) , ∀(m̃ × ñ) ∈ A × B

{max(ζF (m̃) ), ζG(m̃) }, {min(ηF (m̃) ), ηG(m̃) } : λ ∈ U , m̃, ñ ∈ A , B

=
 11

Definition 3.12. For two FFHSS FA and GB , the ”AND” operation on FA and GB
is denoted as

FA ∧ GB = HoverlineA ×B = (K, ã × b̃)

where

(K, m̃ × ñ) = FA (m̃) ∪ GB(ñ) , ∀(m̃ × ñ) ∈ A × B

{min(ζF (m̃) ), ζG(m̃) }, {max(ηF (m̃) ), ηG(m̃) } : λ ∈ U , m̃, ñ ∈ A , B

=

Proposition 3.13. For three FFHSS FA , GB , and HC ,

i. FA ∨ GB = GB ∨ FA ,

ii. FA ∧ GB = GB ∧ FA ,

iii. FA ∨ [GB ∨ HC ] = [FA ∨ GB ] ∨ HC ,

iv. FA ∧ [GB ∧ HC ] = [FA ∧ GB ] ∧ HC ,

v. [FA ∨ GB ]c = FA
c
∧ GcB ,

vi. (FA ∧ GB )c = FA
c
∨ GcB .

Using the Definitions ”OR” and ”AND”, this proposition is easily proved.

Let FA be a FFHSS and u, v ∈ [0, 1], such that u3 + v 3 ≤ 1. Hence, FA is called

(u,v)
(u, v)−constant FFHSS(shown by CA ), if ζF (m̃) = ũ and ηF (m̃) = ṽ ∀m̃.

Definition 3.14. For three FFHSS FA , GB , and HC ,

(0,1)
(i.) If FA = CA and ζ(m̃) = 0, η(m̃) = 1 for all m̃ ∈ A , FA is said to be a relative null
FFHSS with respect to A (shown by F∅ ).
(1,0)
(ii.) If FA = CA and ζ(m̃) = 1, η(m̃) = 0 for all m̃ ∈ A , (FA is called a relative whole
FFHSS with respect to A (shown by FE ).

Obviously, if F = G is taken, F∅ and FE will be null FFHSS and the whole FFHSS.

Proposition 3.15. For the FFHSSs FA , and GB

˜ GB = FA ∪r GB = FA ,
i. FA ∪

˜ GB = FA ∩r GB = FA ,
ii. FA ∩

˜ F∅ = FA ∪r F∅ = FA ,
iii. (FA ∪

˜ F∅ = FA ∩r F∅ = F∅ .
iv. FA ∩

˜ FE = FA ∪r FE = FE ,
v. FA ∪

˜ FE = FA ∩r FE = FE .
vi. FA ∩

Using Definitions 3.1, 3.7, 3.8, 3.9, 3.10, 3.14, this proposition can be easily proved.
12

4 New Coefficients for Fermatean Fuzzy Hypersoft Set

This section is devoted to the notion of KKs and weighted KKs and some aggregation
operators on FFHSS. Further, some fundamental characteristics of new KKs have been
investigated.

Definition 4.1. Choose two FFHSSs FA = {(λi , ζF (m̃k ) (λi ), ηF (m̃k ) (λi )) : λi ∈ U} and
GB = {(λi , ζG(m̃k ) (λi ), ηG(m̃k ) (λi )) : λi ∈ U}. Then,

m X
n 
X
6
6 
IE(FA ) = ζF (m̃k ) (λi ) + ηF (m̃k ) (λi ) , (16)
k=1 i=1
m X n 
X
6
6 
IE(GB ) = ζG(m̃k ) (λi ) + ηG(m̃k ) (λi ) . (17)
k=1 i=1

are called the informational energies of FA , and GB .

Definition 4.2. Choose two FFHSSs FA = {(λi , ζF (m̃k ) (λi ), ηF (m̃k ) (λi )) : λi ∈ U} and
GB = {(λi , ζG(m̃k ) (λi ), ηG(m̃k ) (λi )) : λi ∈ U}. Then,

m X
n 
X
3
3
3
3 
C(FA , GB ) = ζF (m̃k ) (λi ) ∗ ζG(m̃k ) (λi ) + ηF (m̃k ) (λi ) ∗ ηG(m̃k ) (λi ) (18).
k=1 i=1

is called the correlation measure between FA , GB .

Theorem 4.3. Let FA = {(λi , ζF (d˜k ) (λi ), ηF (m̃k ) (λi )) : λi ∈ U}, and GB = {(λi , ζG(m̃k ) (λi ), ηG(m̃k ) (λi )) :
λi ∈ U} be two FFHSSs. Hence,
(1) C(FA , FA ) = IE(FA ),
(2) C(GB , GB ) = IE(GB ).

Definition 4.4. Choose two FFHSSs FA = {(λi , ζF (m̃k ) (λi ), ηF (m̃k ) (λi )) : λi ∈ U} and
GB = {(λi , ζG(m̃k ) (λi ), ηG(m̃k ) (λi )) : λi ∈ U}. Then,

C(FA , GB )
CC(FA , GB ) = p p (19)
IE(FA ) ∗ IE(GB )
Pm Pn 
3
3
3
3 
k=1 i=1 ζF (m̃k ) (λi ) ∗ ζG(m̃k ) (λi ) + ηF (m̃k ) (λi ) ∗ ηG(m̃k ) (λi )
=r r
Pm Pn 
6
6  Pm Pn 
6
6 
k=1 i=1 ζ (λ
F (m̃k ) i ) + η (λ
F (m̃k ) i ) ∗ k=1 i=1 ζG(m̃k ) (λi ) + ηG(m̃k ) (λi )

is called the KK between FA , GB .

Theorem 4.5. Let FA = {(λi , ζF (m̃k ) (λi ), ηF (m̃k ) (λi )) : λi ∈ U} and GB ) = {(λi , ζG(m̃k ) (λi ), ηG(m̃k ) (λi )) :
λi ∈ U} be two FFHSSs. For the KK between them,
(1) 0 ≤ CC(FA , GB ) ≤ 1.
(2) CC((FA , GB ) = CC(GB , GB ),
(3) If FA , GB , that is, for each i, k, ζF (m̃k ) (λi ) = ζG(m̃k ) (λi )) and ηF (m̃k ) (λi ) = ηG(m̃k ) (λi )),
then CC(FA , GB ) = 1.

Proof. It is clear that it is CC(FA , GB ) ≥ 0. So we need to prove the CC(FA , GB ) ≤ 1

 13

part. Using the Equation 18,

m X
n 
X
3
3
3
3 
C(FA , GB ) = ζF (m̃k ) (λi ) ∗ ζG(m̃k ) (λi ) + ηF (m̃k ) (λi ) ∗ ηG(m̃k ) (λi )
k=1 i=1
m 
X
3
3
3
3 
= ζF (m̃k ) (λ1 ) ∗ ζG(m̃k ) (λ1 ) + ηF (m̃k ) (λ1 ) ∗ ηG(m̃k ) (λ1 )
k=1

3
3
3
3 
+ ζF (m̃k ) (λ2 ) ∗ ζG(m̃k ) (λ2 ) + ηF (m̃k ) (λ2 ) ∗ ηG(m̃k ) (λ2 )
+ ···

3
3
3
3 
+ ζF (m̃k ) (λn ) ∗ ζG(m̃k ) (λn ) + ηF (m̃k ) (λn ) ∗ ηG(m̃k ) (λn )

3
3
3
3 
C(FA , GB ) = ζF (m̃1 ) (λ1 ) ∗ ζG(m̃1 ) (λ1 ) + ηF (m̃1 ) (λ1 ) ∗ ηG(m̃1 ) (λ1 )

3
3
3
3 
+ ζF (m̃2 ) (λ1 ) ∗ ζG(m̃2 ) (λ1 ) + ηF (m̃2 ) (λ1 ) ∗ ηG(m̃2 ) (λ1 )
+ ···
!

3
3
3
3 
+ ζF (m̃m ) (λ1 ) ∗ ζG(m̃m ) (λ1 ) + ηF (m̃m ) (λ1 ) ∗ ηG(m̃m ) (λ1 )


3
3
3
3 
+ ζF (m̃1 ) (λ2 ) ∗ ζG(m̃1 ) (λ2 ) + ηF (m̃1 ) (λ2 ) ∗ ηG(m̃1 ) (λ2 )

3
3
3
3 
+ ζF (m̃2 ) (λ1 ) ∗ ζG(m̃2 ) (λ2 ) + ηF (m̃2 ) (λ2 ) ∗ ηG(m̃2 ) (λ2 )
+ ···
!

3
3
3
3 
+ ζF (m̃m ) (λ2 ) ∗ ζG(m̃m ) (λ2 ) + ηF (m̃m ) (λ2 ) ∗ ηG(m̃m ) (λ2 )

+ ···

3
3
3
3 
+ ζF (m̃1 ) (λn ) ∗ ζG(m̃1 ) (λn ) + ηF (m̃1 ) (λn ) ∗ ηG(m̃1 ) (λn )

3
3
3
3 
+ ζF (m̃2 ) (λn ) ∗ ζG(m̃2 ) (λn ) + ηF (m̃2 ) (λn ) ∗ ηG(m̃n ) (λ2 )
+ ···
!

3
3
3
3 
+ ζF (m̃m ) (λn ) ∗ ζG(m̃m ) (λn ) + ηF (m̃m ) (λn ) ∗ ηG(m̃m ) (λn )

m
X
3
3
3
3
= ζF (m̃k ) (λ1 ) ∗ ζG(m̃k ) (λ1 ) + ζF (m̃k ) (λ2 ) ∗ ζG(m̃k ) (λ2 ) + · · ·
k=1
!!

3
3
+ ζF (m̃k ) (λn ) ∗ ζG(m̃k ) (λn )

m
X
3
3
3
3
+ ηF (m̃k ) (λ1 ) ∗ ηG(m̃k ) (λ1 ) + ηF (m̃k ) (λ2 ) ∗ ηG(m̃k ) (λ2 ) + · · ·
k=1
!!

3
3
+ ηF (m̃k ) (λn ) ∗ ηG(m̃k ) (λn )
14

Now we will use the Cauchy-Schwarz inequality,

m 
2
X
6
6
6 
[C(FA , GB )] ≤ ζF (m̃k ) (λ1 ) + ζF (m̃k ) (λ2 ) + · · · + ζF (m̃k ) (λn )
k=1

6
6
6 
+ ηF (m̃k ) (λ1 ) + ηF (m̃k ) (λ2 ) + · · · + ηF (m̃k ) (λn )

6
6
6 
× ζG(m̃k ) (λ1 ) + ζG(m̃k ) (λ2 ) + · · · + ζG(m̃k ) (λn )
!

6
6
6 
+ ηF G(m̃k ) (λ1 ) + ηG(m̃k ) (λ2 ) + · · · + ηG(m̃k ) (λn )

m X
n 
X
6
6 
C(FA , GB )2 ≤ ζF (m̃k ) (λi ) + ηF (m̃k ) (λi )
k=1 i=1
m X n 
X
6
6 
× ζG(m̃k ) (λi ) + ηG(m̃k ) (λi )
k=1 i=1
C(FA , GB )2 ≤ IE(FA ) × IE(GA ).

Hence, CC(FA , GB ) ≤ 1. So 0 ≤ CC(FA , GB ) ≤ 1.

(2) The proof of this item is obvious.
(3) We will use the Equation 19. As we know that ζF (m̃k ) (λi ) = ζG(m̃k ) (λi ) and ηF (m̃k ) (λi ) =
ηG(m̃k ) (λi ), for i, k, we get

C(F ).C(GG )
CC(FA , GB ) = p A p (20)
IE(FA ). GB
Pm Pn 
6
6 
k=1 i=1 ζ F ( m̃ k ) (λi ) + η F (m̃ k ) (λi )
=s   r
Pm Pn
6   6 Pm Pn 
6
6 
k=1 i=1 ζ F (m̃ k ) (λi ) + ηF (m̃dk ) (λ i ) ∗ k=1 i=1 ζ G( m̃ k ) (λi ) + η G( m̃k ) (λ i )

= 1.

Definition 4.6. Let FA = {(λi , ζF (m̃k ) (λi ), ηF (m̃k ) (λi )) : λi ∈ U} and GB = {(λi , ζG(m̃k ) (λi ), ηG(m̃k ) (
lambdai )) : λi ∈ U} be two FFHSSs. Then, their KK is defined as
C(FA , GB )
CCM (FA , GB ) =  (21)
max IE(FA ), IE(GB )
Pm Pn 
3
3
3
3 
k=1 i=1 ζ F ( m̃ k ) (λ i ) ∗ ζ G( m̃k ) (λ i ) + ηF ( m̃k ) (λ i ) ∗ η G( m̃k ) (λ i )
= nP 
6
6  Pm Pn 
6
6 o
m Pn
max k=1 i=1 ζ F (m̃k ) (λ i ) + η F (m̃k ) (λ i ) , k=1 i=1 ζ G(m̃k ) (λ i ) + η G(m̃k ) (λi )

Theorem 4.7. Let FA = {(λi , ζF (m̃k ) (λi ), ηF (m̃k ) (λi )) : λi ∈ U} and GB = {(λi , ζG(m̃k ) (λi ), ηG(m̃k ) (λi )) :
λi ∈ U} be two FFHSSs. For the KK in Equation 21,
(1) 0 ≤ CCM (FA , GB ) ≤ 1.
(2) CCM (FA , GB ) = CCM (GB , FA ),
(3) If FA , GB , that is, for each i, k, ζF (m̃k ) (λi ) = ζG(m̃k ) (λi )) and ηF (m̃k ) (λi ) = ηG(m̃k ) (λi )),
then CCM (FA , GB ) = 1.
The proof of this theorem is similar to Theorem 4.5.
 15

In this day and age, it is critical to examine the weights of FFHSS in practical appli-
cations. The choice may change when the decision-maker assigns different weights for each
alternative in the universe of discourse. As a result, it is critical to plan the weight before
making aPselection. Let Ω = {Ω1 , Ω2 , · · · , Ωm }T be a weight vector for experts such as
m T
Ωk > 0, k=1 PmΩk = 1, and γ = {γ1 , γ2 , · · · , γn } be a weight vector for parameters such
as γi > 0, k=1 γk = 1. By extending Definitions 4.4 and 4.6, we create the weighted
correlation coefficient between FFHSSs in the sections that follow.

Definition 4.8. Let FA = {(λi , ζF (m̃k ) (λi ), ηF (m̃k ) (λi )) : λi ∈ U} and GB = {(λi , ζG(m̃k ) (λi ), ηG(m̃k ) (λi )) :
λi ∈ U} be two FFHSSs.

CW (FA , GB )
CCW (FA , GB ) = p p (2
IEW (FA ). IEW GB
Pm P 
3
3
3
3 
n
k=1 Ω k i=1 γ i ζ F (m̃ k ) (λi ) ∗ ζ G(m̃ k ) (λi ) + η F (m̃ k ) (λi ) ∗ η G(m̃ k ) (λi )
=r P   rP P 
Pm n
6
6 m n
6
6
k=1 Ωk i=1 γi ζF (m̃k ) (λi ) + ηF (m̃k ) (λi ) ∗ k=1 Ωk i=1 γi ζG(m̃k ) (λi ) + ηG(m̃k ) (λi )

is called KK between FA , GB .

Definition 4.9. Let FA = {(λi , ζF (m̃k ) (λi ), ηF (m̃k ) (λi )) : λi ∈ U} and GB = {(λi , ζG(m̃k ) (λi ), ηG(m̃k ) (λi )) :
λi ∈ U} be two FFHSSs.

CW (FA , GB )
CCM W (FA , GB ) =
(23
max IEW FA ∗ IEW (GB
Pm  Pn 
3
3
3
3 
k=1 Ωk i=1 γi ζF (m̃k ) (λi ) ∗ ζG(m̃k ) (λi ) + ηF (m̃k ) (λi ) ∗ ηG(m̃k ) (λi )
= Pm P 
6
6  Pm P 
6
6 
n n
max k=1 Ωk i=1 γi ζF (m̃k ) (λi ) + ηF (m̃k ) (λi ) , k=1 Ωk i=1 γi ζG(m̃k ) (λi ) + ηG(m̃k ) (λi )

is called KK between FA , GB .

Theorem 4.10. Let FA = {(δi , ζF (m̃k ) (λi ), ηF (dm ˜ k ) (λi )) : λi ∈ U} and GB = {(λi , ζG(m̃k ) (λi ), ηG(m̃k ) (λi )) :
λi ∈ U} be two FFHSSs. Then,
(1) 0 ≤ CCM W (FA , GB ) ≤ 1.
(2) CCM W (FA , GB ) = CCM W (GB , FA ),
(3) If FA , GB , that is, for each i, k, ζF (m̃k ) (λi ) = ζG(m̃k ) (λi )) and ηF (m̃k ) (λi ) = ηG(m̃k ) (λi )),
then CCM W (FA , GB ) = 1.

Proof. (1) The inequality CCM W (FA , GB ) ≥ 0 is trivial, and here we only need to prove
16

that CCM W (FA , GB ) ≤ 1. Then,

n n
!
X X 
3
3
3
3 
CM W (FA , GB ) = Ωk γi ζF (m̃k )(λi ) ∗ ζG(m̃k )(λi ) + ηF (m̃k )(λi ) ∗ ηG(m̃k )(λi )
k=1 i=1
n  
X
3
3
3
3 
= Ωk γ1 ζF (m̃k )(λ1 ) ∗ ζG(m̃k )(λ1 ) + ηF (m̃k )(λ1 ) ∗ ηG(m̃k )(λ1 )
k=1
m  
X
3
3
3
3 
+ γ2 ζF (m̃k )(λ2 ) ∗ ζG(m̃k )(λ2 ) + ηF (m̃k )(λ2 ) ∗ ηG(m̃k )(λ2 )
k=1
n  
X
3
3
3
3 
+ ··· + γn ζF (m̃k )(λn ) ∗ ζG(m̃k )(λn ) + ηF (m̃k )(λn ) ∗ ηG(m̃k )(λn )
k=1
(
 
3
3
3
3 
= Ω1 γ 1 ζF (m̃1 )(λ1 ) ∗ ζG(m̃1 )(λ1 ) + ηF (m̃1 )(λ1 ) ∗ ηG(m̃1 )(λ1 )
 
3
3
3
3 
+ Ω2 γ1 ζF (m̃2 )(λ1 ) ∗ ζG(m̃2 )(λ1 ) + ηF (m̃2 )(λ1 ) ∗ ηG(m̃2 )(λ1 )
)
 
3
3
3
3 
+ · · · + Ω m γ1 ζF (m̃m )(λ1 ) ∗ ζG(m̃m )(λ1 ) + ηF (m̃m )(λ1 ) ∗ ηG(m̃m )(λ1 )
(
 
3
3
3
3 
+ Ω1 γn ζF (m̃1 )(λ2 ) ∗ ζG(m̃1 )(λ2 ) + ηF (m̃1 )(λ2 ) ∗ ηG(m̃1 )(λ2 )
 
3
3
3
3 
+ Ω2 γ1 ζF (m̃2 )(λ2 ) ∗ ζG(m̃2 )(λ2 ) + ηF (m̃2 )(λ2 ) ∗ ηG(m̃2 )(λ2 )
)
 
3
3
3
3 
+ · · · + Ω m γ1 ζF (m̃m )(λ2 ) ∗ ζG(m̃m )(λ2 ) + ηF (m̃m )(λ2 ) ∗ ηG(m̃m )(λ2 )
(
 
3
3
3
3 
+ ··· + Ω1 γn ζF (m̃1 )(λn ) ∗ ζG(m̃1 )(λn ) + ηF (ã1 )(δn ) ∗ ηG(ã1 )(δn )
 
3
3
3
3 
+ Ω2 γ1 ζF (m̃2 )(λn ) ∗ ζG(m̃2 )(λn ) + ηF (m̃2 )(λn ) ∗ ηG(m̃2 )(λn ) + ···
)
 
3
3
3
3 
+ Ωm γ1 ζF (m̃m )(λn ) ∗ ζG(m̃m )(λn ) + ηF (m̃m )(λn ) ∗ ηG(m̃m )(λn )
(
√
3 √
3 √
3 √
3 
= Ω1 γ1 ζF (m̃1 )(λ1 ) ∗ γ1 ζG(m̃1 )(λ1 ) + γ1 ηF (m̃1 )(λ1 ) ∗ γ1 ηG(m̃1 )(λ1 )
√
3 √
3 √
3 √
3 
+ Ω2 γ1 ζF (an
˜ 2 )(λ1 ) ∗ γ1 ζG(m̃2 )(λ1 ) + γ1 ηF (am
˜ 2 )(λ1 ) ∗ γ1 ηG(m̃2 )(λ1 )
√
3 √
3 √
3 √
3 
+ · · · + Ωm γ1 ζF (m̃m )(λ1 ) ∗ γ1 ζG(m̃m )(λ1 ) + γ1 ηF (m̃m )(λ1 ) ∗ γ1 ηG(m̃m )(λ1 )
√
3 √
3 √
3 √
3 
+ Ω1 γ2 ζF (m̃1 )(λ2 ) ∗ γ2 ζG(m̃1 )(λ2 ) + γ2 ηF (m̃1 )(λ2 ) ∗ γ2 ηG(m̃1 )(λ2 )
√
3 √
3 √
3 √
3 
+ Ω2 γ2 ζF (m̃2 )(λ2 ) ∗ γ2 ζG(m̃2 )(λ2 ) + γ2 ηF (am ˜ 2 )(λ2 ) ∗ γ 2 ηG(m̃2 )(λ2 )
√
3 √
3 √
3 √
3 
+ · · · + Ωm γ2 ζF (m̃m )(λ2 ) ∗ γ2 ζG(m̃m )(λ2 ) + γ2 ηF (m̃m )(λ2 ) ∗ γ2 ηG(m̃m )(λ2 )
√
3 √
3 √
3 √
3 
+ · · · + Ω1 γn ζF (m̃1 )(λn ) ∗ γn ζG(m̃1 )(λ1 ) + γn ηF (m̃1 )(λn ) ∗ γn ηG(m̃1 )(λn )
√
3 √
3 √
3 √
3 
+ Ω2 γn ζF (m̃2 )(λn ) ∗ γn ζG(m̃2 )(λn ) + γn ηF (m̃2 )(λn ) ∗ γn ηG(m̃2 )(λn )
)
√
3 √
3 √
3 √
3 
+ · · · + Ωm γn ζF (m̃m )(λn ) ∗ γn ζG(m̃m )(λn ) + γn ηF (m̃m )(λn ) ∗ γn ηG(m̃m )(λn )
 17

(
p √
3 p √
3 p √
3 p √
3
= Ω1 γ1 ζF (ñ1 )(λ1 ) ∗ Ω1 γ1 ζG(m̃1 )(λ1 ) + Ω1 γ1 ηF (m̃1 )(λ1 ) ∗ Ω1 γ1 ηG(m̃1 )(λ1 )
p √
3 √
3 p √
3 p √
3
+ Ω2 γ1 ζF (m̃2 )(λ1 ) ∗ γ1 ζG(m̃2 )(λ1 ) + Ω2 γ1 ηF (m̃2 )(λ1 ) ∗ Ω2 γ1 ηG(am ˜ 2 )(λ1 )
p √
3 p √
3 p √
3 p √
3
+ · · · + Ωm γ1 ζF (m̃m )(λ1 ) ∗ Ωm γ1 ζG(m̃m )(λ1 ) + Ωm γ1 ηF (m̃m )(λ1 ) ∗ Ωm γ1 ηG(m̃m )(λ1 )
p √
3 p √
3 p √
3 p √
3 
+ Ω1 γ2 ζF (m̃1 )(λ2 ) ∗ Ω1 γ2 ζG(m̃1 )(λ2 ) + Ω1 γ2 ηF (m̃1 )(λ2 ) ∗ Ω1 γ2 ηG(m̃1 )(λ2 )
p √
3 p √
3 p √
3 p √
3 
+ Ω2 γ2 ζF (m̃2 )(λ2 ) ∗ Ω2 γ2 ζG(m̃2 )(λ2 ) + Ω2 γ2 ηF (m̃2 )(λ2 ) ∗ Ω2 γ2 ηG(m̃2 )(λ2 )
p √
3 p √
3 p √
3 p √
3 
+ · · · + Ωm γ2 ζF (m̃m )(λ2 ) ∗ Ωm γ2 ζG(m̃m )(λ2 ) + Ωm γ2 ηF (m̃m )(λ2 ) ∗ Ωm γ2 ηG(m̃m )(λ2 )
p √
3 p √
3 p √
3 p √
3 
+ · · · + Ω1 γn ζF (m̃1 )(λn ) ∗ Ω1 γn ζG(m̃1 )(λ1 ) + Ω1 γn ηF (m̃1 )(λn ) ∗ Ω1 γn ηG(m̃1 )(λn )
p √
3 p √
3 p √
3 p √
3 
+ Ω2 γn ζF (m̃2 )(λn ) ∗ Ω2 γn ζG(m̃2 )(λn ) + Ω2 γn ηF (m̃2 )(λn ) ∗ Ω2 γn ηG(m̃2 )(λn )
p √
3 p √
3 p √
3 p √
3
+ ··· + Ωm γn ζF (m̃m )(λn ) ∗ Ωm γn ζG(m̃m )(λn ) + Ωm γn ηF (m̃m )(λn ) ∗ Ωm γn ηG(m̃m )(λn )

Using the Cauchy-Schwarz inequality, we get

m m
!
X X 
6
6 
2
CM W (FA , GB ) ≤ Ωk γi ζF (m̃k ) (λi ) + ηF (m̃k ) (λi )
k=1 i=1
m m
!
X X 
6
6 
× Ωk γi ζG(m̃k ) (λi ) + ηG(m̃k ) (λi )
k=1 i=1

CM W (FA , GB )2 ≤ IEM W (FA ).IEM W (GB )

Therefore,
q
CM W (FA , GB ) ≤ IEM W (FA ) ∗ IEM W (GB

and so 0 ≤ CM W (FA , GB ) ≤ 1.
The proof of item (2) is obvious.
(3) From Equation 22, we have
CW (FA , GB )
CCW (FA , GB ) = p p (
IEW (FA ) ∗ IEW (GB
Pm P 
3
3
3
3 
n
k=1 Ωk i=1 γi ζF (m̃k ) (λi ) ∗ ζG(m̃k ) (λi ) + ηF (m̃k ) (λi ) ∗ ηG(m̃k ) (λi )
=r P  r
Pm n
6
6  Pm P
n

6
6
Ω
k=1 k γ
i=1 i ζ (λ
F (m̃k ) i ) + η (λ
F (m̃k ) i ) ∗ k=1 Ωk i=1 γi ζG(m̃k ) (λi ) + ηG(m̃k ) (λi )

Take ζF (m̃k ) (λi ) = ζG(m̃k ) (λi ) and ηF (m̃k ) (λi ) = ηG(m̃k ) (λi ) for all i, k. Hence,
CW (FA , GB )
CCW (FA , GB ) = p p (2
IEW (FA ) ∗ IEW (GB
Pm P 
6
6 
n
k=1 Ωk i=1 γi ζF (m̃k ) (λi ) + ηF (m̃k ) (λi )
=r P  r
Pm n
6
6  Pm P
n

6
6
k=1 Ωk i=1 γi ζF (m̃k ) (λi ) + ηF (m̃k ) (λi ) ∗ k=1 Ωk i=1 γi ζG(m̃k ) (λi ) + ηG(m̃k ) (λi )

= 1. (2
18

Definition 4.11. Let Td˜k = (ζm̃k , ηm̃k ), Tm̃11 = (ζm̃11 , ηm̃11 ), and Tm̃12 = (ζm̃12 , ηm̃12 ) be
three FFHSNs and α be a positive real number; by algeabric norms,we have
q 
3
(1) Tm̃11 ⊞ Tm̃12 = 3 ζm̃ + ζ 3 − ζ 3 ζ3 , η η
11 m̃12 m̃11 m̃12 m̃ 11 m̃12 ,

 q 
3
(2) Tm̃11 ⊠ Tm̃12 = ζm̃11 ζm̃12 3 ηm̃ + η 3 − η 3 η 3 ,
11 m̃12 m̃11 m̃12

q
α 
3 α
(3) αTm̃k = 3
1 − 1 − ζm̃ k
, ηm̃k

 q
α 
α α 3
(4) Tm̃ k
= ζ m̃k , 3
1 − 1 − η m̃k ,

Based on the foregoing rules for the collection of FFHSNs, certain averaging and geo-
metric aggregation operations for FFHSSs have been established.

Definition 4.12. Let Td˜ij = (ζm̃ij , ηm̃ij ) be a FFHSN, Ωi and γj be the weighted vectors
Pn Pn
with Ωi > 0, i=1 Ωi = 1, γj > 0, and i=1 γi = [Link], FFHS weighted averaging
operator is characterized by F F HSW A : ∆n → ∆, which is given as follows:

F F HSW A(Tm̃11 , Tm̃12 , · · · , Tm̃nm ) = ⊞m n

j=1 γj ⊞i=1 Ωi Tm̃ij
v !γj 
Ωj γj Y
u m n 
! m n
u Y Y Y
Ωj
= t 3
1− 1 − ζ3 ,
m̃ij ηm̃ .
ij
j=1 i=1 j=1 i=1

Definition 4.13. Let Td˜ij = (ζd˜ij , ηd˜ij ) be a FFHSN, Ωi and γj be the weighted vectors
Pn Pn
with Ωi > 0, i=1 Ωi = 1, γj > 0, and i=1 γi = 1. Then, FFHS weighted geometric
operator is defined as F F HSW G : ∆n → ∆, which is defined as follows:

F F HSW G(Tm̃11 , Tm̃12 , · · · , Tm̃nm ) = ⊠m n

j=1 γj ⊠i=1 Ωi Tm̃ij
 !γj v ! 
Ym Yn

Ωj
u
u m
Y n 
Y Ωj γj
=  ζm̃ij 3
,t 1− 3
1 − ηm̃ 
ij
j=1 i=1 j=1 i=1

5 New Approach
In this part, we will continue the TOPSIS approach for FFHSS information based on KKs
to construct a framework for solving DM concerns. The TOPSIS approach was created and
applied by Yoon Hwang [48] to promote the order of assessment components of positive(PIS)
and negative ideal solutions(NIS) for DM challenges. Using the TOPSIS technique, we will
be able to find the best potential options with the shortest and biggest distances to the PIS
and NIS, respectively. By using rankings, the TOPSIS approach assures that the correlation
measure can discriminate between positive and negative ideals. In general, researchers use
the TOPSIS approach to determine proximity coefficients, unique distance forms, and sim-
ilar measurements. By using rankings, the TOPSIS approach assures that the correlation
measure may be utilized to discriminate between positive and negative ideals. In general,
researchers use the TOPSIS approach to determine proximity coefficients, unique distance
forms, and similar measurements. Because the correlation measure preserves the linear rela-
tionship between the elements investigated, the TOPSIS approach using KKs is preferable
for identifying closeness coefficients rather than distance and similarity measurements. A
 19

TOPSIS approach for selecting the best choice is shown by leveraging the created KK.

Scenario:
Let’s consider a set of ”s” alternatives such as A = {A1 , A2 , · · · , As } for the evaluation
of the E P = {E1 , E2 , · · · , En } expert team with Ω = (Ω1 , Ω2 , · · · , Ωn )T weights providing
n
Ωi > 0, i=1 Ωi = 1. Let D = {d1 , d2 , · · · , dm } specified as a set of attributes. Let
T = {(t1p × t2p × · · · × dmp ), for all p ∈ {1, 2, · · · , t}} be a collection of sub-attirbutes with
Pt
γ = (γ1p , γ2p , · · · , γmp )T weights satisfying the conditions
P
= 1 γp > 0, p=1 γp = 1. The
elements in the collection of sub-attributes are multi-valued; for the sake of convenience,
the elements of T can be expressed as T = {d˜∂ : ∂ ∈ {1, 2, cdots, k}}. The team of experts
{Ei : i, 1, 2, · · · , n} evaluate the alternatives {A(z) : z = 1, 2, · · · , s} based on the desired
sub-attributes of the considered parameters {d˜∂ : ∂ = 1, 2, · · · , k} given in the form of FFH-
(z) (z) (z) (z) (z) (z) (z)
SNs such as (Td˜ )n×∂ = (ζd∂
∂
˜ )n×∂ , where 0 ≤ ζd˜ , ηd˜ ≤ 1 and ζd˜3 +ηd˜3 ≤ 1 for all i, j.
˜ , ηd∂ ij ij ij ij

Algorithm:

Step 1. Create a matrix in the form of FFHSNs for each alternative {A(z) : z =
1, 2, · · · , s} by utilizing sub-attributes of the supplied attributes such as:

 (z) (z) (z) (z) (z) (z)


(ζ ˜ , ηd˜ ) (ζd˜ , ηd˜ ) · · · ζd˜ , ηd˜
 d(z)11 11 12 12 1∂ 1∂
(ζ ˜ , η (z)
˜ ) (ζ
(z)
, η
(z)
) ···
(z) (z) 
ζd˜ , ηd˜ 
d˜22 d˜22
(A (z)
, T )n×δ =  d21 . d21 2∂ 2∂ 
(27)

.. .
.. .. .. 

 . . 

(z) (z) (z) (z) (z) (z)
(ζd˜ , ηd˜ ) (ζd˜ , ηd˜ ) · · · ζd˜ , ηd˜
11 11 n1 n1 n∂ n∂

Step 2: Normalize the collective information decision matrix by using the normalization
procedure to turn the rating value of the cost-type parameters into benefit-type parameters:

  
 T ˜c = η (z) , ζ (z) , for cost-type parameter
hij = dij  d˜ij d˜ij 
 T˜ = ζ ,η(z) (z)
dij d˜ d˜
, for benefit-type parameter
ij ij

(z) (z)
Step 3: Construct the weighted decision matirx for each alternative (A ) = (T d˜ij )n×∂ ,
where

s !
(z)
 Ωi γj  Ωi γj  (z) 
(z) (z)
T d˜ij = γj Ωi Td˜ = 3
1− 1− ζd3˜ , ηd˜ij = ζ d˜ij , η d˜ (28)
ij ij ij

(z) (z)
Step 4: Find the indices hij = argmaxz {θij } and gij = argminz {θij } for each expert
Ei and sub-attribute d˜j from correlation coefficient matrices and determine the PIA and
NIA based on indices as follows:
hij h
L+ = (ζd+ + ij
˜ , ηd˜ )n×∂ = (ζ d˜ij , η d˜ ) (29)
ij ij ij
gij g
L− = (ζd− −
˜ , ηd˜ )n×∂ = (ζ d˜ij , η d˜ij ) (30)
ij ij ij

(z)
Step 5: Calculate the KK between each alternative of weighted decision matrices A
and PIA L+ :
20

(z)
p(z) = CC(A , L+ ) (31)
 
Pm Pn (z) (z)
j=1 i=1 ∗ + ∗
ζd˜ ζd+
˜ij ηd˜ ηd+
˜ij
ij ij
= s  s
3  3  P  3  3 
3
Pm Pn (z) (z) 3 m Pn + +
j=1 i=1 ζ d˜
+ ηd˜
. j=1 i=1 ζd˜
+ η d˜
ij ij ij ij

Step 6: Compute the correlation coefficient between each alternative of weighted deci-
(z)
sion matrices A and NIA L− :

(z)
q (z) = CC(A , L− ) (32)
 
Pm Pn (z) (z)
j=1 i=1 ∗ +ζd˜∗ ζd−
˜ij ηd˜ ηd−
˜ij
ij ij
= s   s 
3  3 3  3 
3
Pm Pn (z) (z) 3
Pm Pn − −
j=1 i=1 ζd˜ + ηd˜ . j=1 i=1 ζd˜ + ηd˜
ij ij ij ij

Step 7: Calculate the closeness coefficient for each alternative:

(z)
H(A , L− )
R(z) = (z) (z)
(33)
H(A , L+ ) + H(A , L− )

where
(z)
H(A , L− ) = 1 − q (z) ,
(z)
H(A , L+ ) = 1 − p(z) .

Step 8: Select the alternative with a maximum value of the closeness coefficient.

Step 9: Analyze the ranking of the alternatives.

6 Numerical Examples
This section has been written to verify the suggested approach. The material for the femoral
component of the hip joint prosthesis will be chosen following the procedure described in the
preceding section. There are several approaches in the literature for selecting this biomedical
material.

The femoral head and the acetabulum are the two main components of the hip joint,
which is a crucial load-bearing joint in the human body. The femoral head is what at-
taches to the acetabulum, a socket in the pelvic bone, to create the hip joint. The femoral
head can move inside and outward, anterior and backward, and in circles within the ac-
etabulum (Fig. 4). The hip joint’s primary function is load carrying, therefore it has to
have a suitable range of motion and stability. Hip arthroplasty is a surgical treatment used
to replace or rejuvenate a hip joint that has been badly damaged or calcified (osteoarthritis).

The best course of treatment for significant discomfort, mobility restrictions, and short-
ness that interfere with everyday activities is hip replacement. The success of applying hip
 21

Figure 2: Decision-making process

Figure 3: Flow chart

22

Figure 4: Healthy Hip [29]

prostheses is increased by using implants with the right materials and design elements. Nu-
merous prostheses have been created in the present day with various construction materials
and characteristics. The prosthesis should be easy to manufacture, affordable, dependable,
and long-lasting thanks to the design and material characteristics of the implant that has
been chosen. The choice of materials for hip replacement prostheses is problematic since
the design calls for a number of distinctive key qualities that are challenging to develop in
a single material.

An ideal biomaterial is expected to have extremely high biocompatibility, meaning there

won’t be any negative tissue reactions when it’s implanted within the human body. In addi-
tion, it must have bone-like density, high mechanical power, low elastic modulus, and good
corrosion and fatigue resistance.

The femoral component, the acetabular cup, and the acetabular interface are the three
primary components of a hip prosthesis. The femoral component, the acetabular cup, and
the acetabular interface are the three primary components of a hip prosthesis. To replace the
native femoral head, a robust metal pin is placed into the hollowed-out shaft of the femur.
To replace the native femoral head, a robust metal pin is placed into the hollowed-out shaft
of the femur. The acetabulum is filled by an acetabular cup, a soft polymer molding linked
to the ilium. The acetabular interface is positioned between the femoral component and the
acetabular cup and can be made of a variety of materials to reduce frictional wear debris.
The hip joint prosthesis is shown in Fig. 5 in its normal shape and location.

Various stresses and effects are applied to biomaterials in various human body compo-
nents. For instance, daily activities exert strains of 4 MPa on bones and 40–80 MPa on
tendons. When doing actions like jumping, the strain on a hip joint can increase to up to
ten times its weight from the typical three times. Throughout the course of the day, actions
including standing, jogging and sitting cause these stressors in the body repeatedly. Bioma-
terials may become worn down, break, or deform plastically due to these repeated motions.

It has been experimentally measured that the femoral head is loaded up to 3.5 times the
 23

Figure 5: Hip joint prosthesis [30]

body weight (80 kg body weight) during walking, showing that total hip prostheses must
be sufficiently resistant to these loads. At the same time, these prostheses must be resistant
to abrasion caused by friction in the joint. Today, in a hip prosthesis; Vitallium (Co-Cr-Mo
alloy), stainless steel, high-density polyethylene, polymethylmethacrylate, and titanium and
titanium alloys are used.

Let’s assume that there are four orthopedists in the group that will decide on the selec-
tion of biomaterials(O = {O1 , O2 , O3 , O4 }).

Orthopedists with weights (0.2, 0.3, 0.4, 0.1)T , including A = {A1 , A2 , A3 , A4 }, (vitallium(Co-
Cr-Mo), stainless steel, high-density polyethylene, polymethylmethacrylate, titanium and
titanium alloys) the set of biomedical materials, evaluate the grades of these material types.
The group of orthopedists decides the criteria(attirbute) for the choice of biomedical mate-
rials as: L = {ℓ1 , ℓ2 , ℓ3 } (strength, resistance, tolerance). The sub-attributes corresponding
to these attributes are: ℓ1 = {d11 , d12 } (Tensile strength, Fatigue strength), ℓ2 = {d21 , d22 }
(corrosion resistance, relative wear resistance), ℓ3 = {d31 , d32 } (Tissue tolerance, elasticity).
′
Let T = ℓ1 × ℓ2 × ℓ3 be a set of sub-attributes:

′
T = ℓ1 × ℓ2 × ℓ3 = {d11 , d12 } × {d21 , d22 } × {d31 , d32 }
(
= (d11 , d21 , d31 ), (d11 , d21 , d32 ), (d11 , d22 , d31 ), (d11 , d22 , d32 ),
)
(d12 , d21 , d31 ), (d12 , d21 , d32 ), (d12 , d22 , d31 ), (d12 , d22 , d32 )

′
Let T = {d˜1 , d˜2 , d˜3 , d˜4 , d˜5 , d˜6 , d˜7 , d˜8 } be a set of all multi-sub-attributes with weights
(0.12, 0.18, 0.1, 0.15, 0.05, 0.22, 0.08)T . Each orthopedist will evaluate the ratings of biomed-
ical materials in the form of FFHSNs for each sub-attribute of the considered parame-
ters(Tables 1-4). The developed method to find the best alternative is as follows:
24

Step 1: Create decision matrices for each alternative under defined multi-sub-attributes
based on each decision-FFHSN maker’s rating.

Step 2: Because all of the criteria are of beneficial kinds, they must be normalized.

Step 3: Using Equation 28 from Tables 5-8, create a weighted decision matrix for each
(z) (z)
alternative A = (Lij )n×∂ .

Table 1: Decision Matrix for A(1)

A(1) d˜1 d˜2 d˜3 d˜4 d˜5 d˜6 d˜7 d˜8
O1 (0.33, 0.75) (0.72, 0.29) (0.64, 0.75) (0.59, 0.32) (0.25, 0.35) (0.46, 0.54) (0.52, 0.76) (0.88, 0.32)
O2 (0.59, 0.73) (0.37, 0.64) (0.25, 0.49) (0.92, 0.27) (0.34, 0.82) (0.27, 0.46) (0.73, 0.55) (0.44, 0.65)
O3 (0.75, 0.35) (0.28, 0.61) (0.17, 0.66) (0.34, 0.48) (0.46, 0.69) (0.87, 0.47) (0.61, 0.78) (0.28, 0.53)
O4 (0.83, 0.45) (0.29, 0.84) (0.25, 0.47) (0.46, 0.68) (0.63, 0.58) (0.55, 0.69) (0.47, 0.56) (0.83, 0.37)

Table 2: Decision Matrix for A(2)

A(2) d˜1 d˜2 d˜3 d˜4 d˜5 d˜6 d˜7 d˜8
O1 (0.69, 0.64) (0.32, 0.46) (0.67, 0.55) (0.33, 0.82) (0.58, 0.52) (0.46, 0.62) (0.73, 0.61) (0.47, 0.81)
O2 (0.83, 0.55) (0.73, 0.46) (0.93, 0.19) (0.75, 0.33) (0.43, 0.65) (0.95, 0.16) (0.28, 0.67) (0.31, 0.82)
O3 (0.36, 0.72) (0.48, 0.57) (0.42, 0.78) (0.35, 0.65) (0.62, 0.65) (0.37, 0.57) (0.91, 0.18) (0.68, 0.23)
O4 (0.53, 0.54) (0.69, 0.58) (0.87, 0.25) (0.86, 0.48) (0.93, 0.25) (0.27, 0.44) (0.53, 0.66) (0.67, 0.35)

Table 3: Decision Matrix for A(3)

A(3) d˜1 d˜2 d˜3 d˜4 d˜5 d˜6 d˜7 d˜8
O1 (0.51, 0.71) (0.83, 0.46) (0.68, 0.45) (0.46, 0.36) (0.48, 0.82) (0.28, 0.38) (0.83, 0.44) (0.75, 0.54)
O2 (0.79, 0.52) (0.77, 0.34) (0.86, 0.22) (0.55, 0.24) (0.53, 0.68) (0.72, 0.46) (0.66, 0.59) (0.61, 0.42)
O3 (0.63, 0.78) (0.45, 0.56) (0.61, 0.63) (0.65, 0.48) (0.69, 0.54) (0.77, 0.46) (0.54, 0.77) (0.46, 0.52)
O4 (0.48, 0.67) (0.89, 0.31) (0.35, 0.54) (0.56, 0.68) (0.36, 0.55) (0.77, 0.51) (0.77, 0.39) (0.34, 0.38)

Table 4: Decision Matrix for A(4)

A(4) d˜1 d˜2 d˜3 d˜4 d˜5 d˜6 d˜7 d˜8
O1 (0.50, 0.69) (0.79, 0.54) (0.76, 0.45) (0.43, 0.36) (0.48, 0.92) (0.26, 0.46) (0.81, 0.44) (0.72, 0.52)
O2 (0.77, 0.45) (0.63, 0.48) (0.83, 0.55) (0.57, 0.26) (0.53, 0.69) (0.70, 0.46) (0.69, 0.63) (0.61, 0.45)
O3 (0.56, 0.45) (0.44, 0.77) (0.52, 0.67) (0.35, 0.49) (0.72, 0.61) (0.68, 0.55) (0.41, 0.88) (0.58, 0.18)
O4 (0.43, 0.67) (0.18, 0.34) (0.74, 0.45) (0.56, 0.78) (0.33, 0.54) (0.77, 0.34) (0.33, 0.58) (0.27, 0.55)

Step 4: Determine the PIA and NIA based on indices by using Equations 29 and 30:
 25

 

 (0.5987, 0.8799), (0.5295, 0.9167), (0.6019, 0.9426), (0.6249, 0.9784), (0.8194, 0.9704),

(0.4874, 0.9678), (0.7787, 0.9489), (0.6219, 0.9572) (0.7152, 0.9726), (0.5909, 0.9468),

 

 

(0.6875, 0.9436), (0.7567, 0.9289), (0.8977, 0.9710), (0.6459, 0.9533), (0.7965, 0.9899),

 

+
L = (0.6948, 0.9903), (0.7345, 0.9428), (0.6122, 0.9313), (0.7170, 0.9698), (0.8427, 0.9583),
(0.4957, 0.9115), (0.7639, 0.9388), (0.7709, 0.9617), (0.4783, 0.8874) (0.6523, 0.9905),

 

 

(0.4520, 0.9867), (0.6987, 0.9861), (0.7468, 0.9556), (0.8575, 0.9914), (0.4431, 0.9876),

 

 

(0.7784, 0.9961), (0.6447, 0.9808)
 

 

 (0.7465, 0.9753), (0.6427, 0.9645), (0.7802, 0.9726), (0.6937, 0.9476), (0.8895, 0.9786),

(0.5778, 0.9629), (0.8219, 0.9648), (0.7725, 0.7762) (0.7359, 0.9576), (0.6968, 0.9681),

 

 

(0.8964, 0.9123), (0.8946, 0.9133), (0.8895, 0.9269), (0.8827, 0.9687), (0.4462, 0.9036),

 

L− = (0.8471, 0.9445), (0.8578, 0.9075), (0.7197, 0.9313), (0.8892, 0.9426), (0.8167, 0.9365),
(0.8026, 0.9743), (0.6812, 0.9416), (0.8547, 0.9702), (0.8967, 0.9874) (0.9053, 0.9315),

 

 

(0.7285, 0.9862), (0.5028, 0.9748), (0.7438, 0.9859), (0.6618, 0.9772), (0.7614, 0.9839),

 

 

(0.6575, 0.9576), (0.7855, 0.9822)
 

(z)
Step 5: Compute the KK between A and PIA L+ by using Equation 31, given
p = 0.97724, p = 0.96823, p = 0.96182, p(4) = 0.96221.
(1) (2) (3)

(z)
Step 6: Compute the KK between A and NIA L− by using Equation 31, given
q = 0.95651, q = 0.96107, q = 0.96463, q (4) = 0.96501.
(1) (2) (3)

(1)
Table 5: Weighted Decision Matrix for A
A(1) d̃1 d̃2 d̃3 d̃4 d̃5 d̃6 d̃7 d̃8
O1 (0.716, 0.983) (0.548, 0.957) (0.6681, 0.9852) (0.664, 0.948) (0.815, 0.909) (0.502, 0.987) (0.8381, 0.9743) (0.6719, 0.9726)
O2 (0.7137, 0.9782) (0.7308, 0.9827) (0.7895, 0.9927) (0.9190, 0.9311) (0.8704, 0.9617) (0.5674, 0.9314) (0.8312, 0.9735) (0.8272, 0.9849)
O3 (0.7534, 0.9384) (0.6322, 0.9412) (0.7362, 0.9648) (0.7469, 0.9345) (0.9019, 0.9768) (0.4868, 0.9258) (0.8132, 0.9592) (0.7791, 0.9647)
O4 (0.6382, 0.9817) (0.5114, 0.9696) (0.6835, 0.9901) (0.6355, 0.9793) (0.8547, 0.9695) (0.4874, 0.9873) (0.7686, 0.9864) (0.6434, 0.9879)

(2)
Table 6: Weighted Decision Matrix for A
A(2) d̃1 d̃2 d̃3 d̃4 d̃5 d̃6 d̃7 d̃8
O1 (0.6786, 0.9789) (0.6278, 0.9786) (0.7891, 0.9794) (0.6749, 0.9746) (0.9011, 0.9885) (0.5924, 0.9678) (0.8137, 0.9768) (0.7892, 0.9825)
O2 (0.7609, 0.9742) (0.6069, 0.9493) (0.7108, 0.9504) (0.9111, 0.9294) (0.4599, 0.9206) (0.8029, 0.9879) (0.7046, 0.9879) (0.6902, 0.9731)
O3 (0.7912, 0.9495) (0.7408, 0.9518) (0.7785, 0.9902) (0.7315, 0.9414) (0.6377, 0.9189) (0.8824, 0.9726) (0.7686, 0.9392) (0.9104, 0.9388)
O4 (0.7316, 0.9819) (0.4389, 0.9839) (0.5588, 0.9812) (0.6792, 0.9789) (0.9899, 0.9891) (0.4408, 0.9769) (0.7613, 0.9913) (0.6706, 0.4335)

(3)
Table 7: Weighted Decision Matrix for A
A(3) d̃1 d̃2 d̃3 d̃4 d̃5 d̃6 d̃7 d̃8
O1 (0.7499, 0.9814) (0.6099, 0.9674) (0.7902, 0.9826) (0.6708, 0.9730) (0.8515, 0.9901) (0.5978, 0.9708) (0.8112, 0.9845) (0.7625, 0.9826)
O2 (0.7613, 0.9687) (0.6095, 0.9491) (0.7106, 0.9469) (0.8984, 0.9903) (0.4678, 0.9263) (0.6102, 0.9671) (0.8107, 0.9905) (0.7106, 0.9866)
O3 (0.8687, 0.9175) (0.7005, 0.9326) (0.9116, 0.9592) (0.7407, 0.9510) (0.8606, 0.9718) (0.6909, 0.9418) (0.8489, 0.9718) (0.8868, 0.9488)
O4 (0.7316, 0.9901) (0.5143, 0.9699) (0.7481, 0.9904) (0.6105, 0.9830) (0.8519, 0.9924) (0.6102, 0.9805) (0.7843, 0.9827) (0.7612, 0.9804)

Step 7: Compute the closeness coefficient by using Equation 33, R(1) = 0.77498,
R = 0.55478, R(3) = 0.43854, R(4) = 0.47487.
(2)

Step 8: Choose the alternative with maximum closeness coefficient R(1) = 0.77498, so
(1)
A is the best alternative.
26

(4)
Table 8: Weighted Decision Matrix for A
A(4) d̃1 d̃2 d̃3 d̃4 d̃5 d̃6 d̃7 d̃8
O1 (0.6905, 0.9905) (0.6702, 0.9547) (0.7131, 0.9804) (0.7104, 0.9624) (0.8124, 0.9790) (0.5108, 0.9615) (0.8107, 0.9784) (0.7291, 0.9795)
O2 (0.7608, 0.9699) (0.6127, 0.9526) (0.9103, 0.9343) (0.7801, 0.9291) (0.9114, 0.9786) (0.6494, 0.9535) (0.9116, 0.9596) (0.8617, 0.9545)
O3 (0.8716, 0.9398) (0.7409, 0.9467) (0.8905, 0.9299) (0.8497, 0.9638) (0.6501, 0.9465) (0.8620, 0.9704) (0.9007, 0.9612) (0.8514, 0.9772)
O4 (0.7313, 0.9904) (0.4424, 0.9718) (0.7216, 0.9913) (0.6721, 0.9812) (0.8520, 0.9956) (0.6605, 0.9560) (0.7684, 0.9954) (0.6909, 0.9912)

Step 9: Analyzing the ranking of the alternatives, we can see R(1) > R(2) > R(4) >
(1) (2) (4) (3)
R(3) , so the ranking of the alternatives is A > A > A > A .

7 Discussion
7.1 Comparison
We may use Zadeh’s [51] approach to process the MD of the attributes, however, this method
cannot handle the NMD and sub-attributes of the considered parameters. Zhang et al. [52]
employed MD and NMD to deal with uncertainty, although these theories have limitations,
such as when the sum of MD and NMD exceeds one, these theories are unable to handle
the situation. Yager [46] developed the PFS to solve these concerns, but it is incapable of
dealing with the parametric values of the alternatives. The FSS was created by Maji et al.
[26] to handle various parameterizations.
The FSS is unaware of the NMD of the alternative’s qualities and is only concerned
with the MD of the attributes. In contrast, the MD and NMD are used in our proposed
PFHSS to control uncertainty. Maji et al. [27] proposed using MD and NMD of features
with parameterization to account for uncertainty in the IFSS. The IFSS cannot deal with
problems when the sum of MD and NMD is more than one. In contrast to the IFSS, Peng
et al. [31] developed the PFSS to effectively manage ambiguity. All of the aforementioned
theories fail to manage situations where qualities contain sub-attributes.
The FSS does not know the NMD of the characteristics of the alternative, and it solely
deals with the MD of the attributes. In contrast, our proposed PFHSS manages uncertainty
by employing the MD and NMD. Maji et al. [27] suggested the IFSS account for uncertainty
by employing MD and NMD of characteristics with parameterization. The IFSS is unable
to tackle issues where the total of MD and NMD is more than one. In comparison to IFSS,
Peng et al. [31] created the PFSS to handle ambiguity properly. When qualities contain sub-
attributes, all of the aforementioned theories fail to manage the circumstance. Zulqarnain
et al.[54] used the IFHSS environment to propose the TOPSIS technique for dealing with
uncertain scenarios using MD and NMD in which the total of the researched parameters’ sub-
attributes cannot exceed one. When the sum of MD and NMD of sub-attributes exceeds
one, as in ζF (m̃) (λ) + ηF (m̃) (λ) ≥ 1 IFHSS cannot handle the situation. To get over the
above-mentioned limits, we transformed the IFHSS to PFHSS by modifying the condition
ζF (m̃) (λ) + ηF (m̃) (λ) ≤ 1 to ζF (m̃) (λ)2 + ηF (m̃) (λ)2 ≤ 1. Instead, as demonstrated in Table
51, the technique we proposed is a complex strategy that can cope with alternatives with
varying sub-attribute information. On the contrary, the developed method in this work
addresses uncertainty by employing the MD and NMD of various sub-attributes. As a
result, the suggested approach outperforms existing methodologies and, without a doubt,
generates superior results for decision-makers during the decision-making process.
According to the information given in Table 9, we can briefly explain the advantages of
the methods:

FS [51] handles uncertainty using fuzzy intervals. IFS[52] handles uncertainty using MD
and NMD. PFS([46]) deals with more uncertainty using MD and NMD than IFS. FSS[26]
 27

Table 9: Comparative Analysis

Set Truthiness Falsity Attributes Sub-attributes
FS [51] YES NO YES NO
IFS[52] YES YES YES NO
PFS([46]) YES YES YES NO
FSS[26] YES NO YES NO
IFSS[27] YES YES YES NO
PFSS[31] YES YES YES NO
IFHSS[54] YES YES YES YES
PFHSS[55] YES YES YES YES
FFHSS(Our method) YES YES YES YES

handle uncertainty using fuzzy range with their parameterization. IFSS[27] deals with more
uncertainty using MD and NMD with their parameterizations. PFSS [31] deals with more
uncertainty than IFSS. IFHSS[54] handles the uncertainty of multiple sub-attributes using
MD and NMD. PFHSS[55] deals with more uncertainty than IFHSS. FFHSS, by definition,
deals with more uncertainty than PFHSS and IFHSS.

7.2 Superiority of the new approach

The FFS is a generalization of the classical set, which consists of the FS, IFS, and PFS. PFS,
as one of the most common extensions, is differentiated by the degrees of membership and
nonmembership satisfaction with regard to the criteria, where their total squares are equal
to or less than 1. In some cases, the decision-maker may offer the degree of membership and
nonmembership of a specific feature in such a way that the sum of the square is greater than
1. As a result, the PFS does not handle this instance effectively. To address this limitation,
FFS theory is one of the more extensive ones that can handle not just partial information
but also uncertain information and inconsistent information, both of which are commonly
present in practical settings. As a result, actual scientific and technical applications are
better suited for decision-making using Fermatean fuzzy information.

Making judgments regarding real-world situations, or finding solutions to them, is chal-

lenging and complex. Therefore, it is crucial to reduce uncertainty while making a decision
in order to select the best option or alternatives. The connections between the inputs must
also be handled effectively for decision-making to be most beneficial. We put out a fresh
approach that combines the benefits of FFS and KK to demonstrate the presence of the
association between the variables.

A combination of interval-valued fermatean fuzzy sets with Fermatean hesitant fuzzy

components in the form of interval values is known as an interval-valued fermatean hesitant
set. Since Fermatean hesitant fuzzy sets are effective instruments for representing more
complex, ambiguous, and hazy information, interval-valued Fermatean hesitant fuzzy sets
are expansions of these sets. IVFHFS is explained and used in the process of group decision-
making. Since membership and non-membership are taken into account concurrently, and
IVFHFS offers a greater feasible zone between 0 and 1 when decision-makers express their
agreement and disagreement, it can provide decision-makers more freedom in analyzing un-
clear and fuzzy information.

A number of academics have developed an approach that uses the KK for PFSs, as seen
from the previous studies. As was already established, some scenarios cannot be captured
28

by PFSs, hence the related algorithm may not produce the desired outcomes.

A specific case of the KK for FFSs is the KK for PFSs. The suggested KK is therefore
more applicable to a wider range of situations than the current ones, making it better suited
to real-world problem solving.

We concluded from this research and comparative analysis that the outcomes of our
method are more universal than those of existing procedures. However, when compared
to previous decision-making methodologies, the decision-making process incorporates more
information to deal with data ambiguity. Furthermore, many FS hybrid structures have
evolved into FFHSS special instances, and certain acceptable criteria have been added.
Among these, object-related information may be conveyed more correctly and experimen-
tally, making it a useful tool for combining erroneous and ambiguous information in the
decision-making process. As a result, our suggested solution is more effective, versatile,
simple, and superior to previous hybrid fuzzy set architectures.

8 Conclusion
The concept of FFHSS is used to handle problems with insufficient information, ambiguity,
and inconsistency by taking the MD and ND of the subattributes of the examined attributes
into account. We created the KK and WKK for FFHSS and exhibited their favorable proper-
ties. Similarly, an enhanced TOPSIS technique is proposed based on the created correlation
by taking into account the attribute set with its matching sub-attributes and decision-
makers. We created correlation indices to discover PIA and NIA. The proximity coefficients
were constructed based on the well-established TOPSIS approach for rating alternatives. A
numerical demonstration of how to solve the MAGDM issue using the suggested TOPSIS
approach has been provided. Furthermore, a comparison study was performed to validate
the efficacy and presentation of the provided strategy. Finally, based on the findings, it is
possible to infer that the suggested approach provides more stability and practicability for
decision-makers during the decision-making process. Future studies will focus on providing
ideas to various operators in the FFHSS environment about decision-making problems. In
a given setting, several different structures, such as topological structures, algebraic struc-
tures, and ordered structures, can be built and examined. This research paper has pragmatic
bounds and may be quite useful in real-world aspects such as sickness diagnosis, pattern de-
tection, and economics. We are confident that this publication will open new doors for
scholars in this subject.

References
[1] Atanassov K.T. Intuitionistic fuzzy sets. Fuzzy Sets and Systems 1986; 20 (1): 87-96.
[Link]/10.1016/S0165-0114(86)80034-3

[2] Chiang D.A., Lin N.P. Correlation of fuzzy sets. Fuzzy Sets and Systems 1999; 202 (2):
221-226. [Link]/10.1016/S0165-0114(97)00127-9

[3] Donghai L., Yuanyuan L., Xiaohong C. Fermatean fuzzy linguistic set and its applica-
tion in multicriteria decision making. International Journal of Intelligent Systems 2019;
34 (3): 878–894. doi: 10.1002/int.22079.

[4] Donghai L, Yuanyuan L, Lizhen W. Distance measure for Fermatean fuzzy linguistic
term sets based on linguistic scale function: An illustration of the TODIM and TOP-
 29

SIS methods. International Journal of Intelligent Systems 2019; 34 (1): 2807–2834.

doi:10.1002/int.22162.

[5] Drugwatch, Hip Replacement, [Link]

[6] Ejegwa, P.A. Distance and similarity measures for Pythagorean fuzzy sets. Granul.
Comput. 2018; 5, 225–238. [Link]

[7] Ejegwa, P.A., Onasanya BO. Improved intuitionistic fuzzy composite relation and its
application to medical diagnostic process, Notes on Intuitionistic Fuzzy Sets 2018; 25,
43–58. [Link]

[8] Garg H. A New Generalized Pythagorean Fuzzy Information Aggregation Using Ein-
stein Operations and Its Application to Decision Making. Int J Intell Syst, 2016; 31:886–
920.

[9] Garg H. Some series of intuitionsitic averaging aggregation operators. SpringerPlus

2016; 5(1), 999, DOI: 10.1186/s40064-016-2591-9

[10] Garg H. Shahzadi G., Akram M. Decision-Making Analysis Based on Fermatean Fuzzy
Yager Aggregation Operators with Application in COVID-19 Testing Facility. Mathe-
matical Problems in Engineering 2020; Article ID 7279027. doi:10.1155/2020/7279027.

[11] Khan, A.A., Ashraf, S., Abdullah, S., Qiyas, M., Luo, J., Khan, S.U. Pythagorean
Fuzzy Dombi Aggregation Operators and Their Application in Decision Support Sys-
tem. Symmetry 2019: 11, 383. [Link]

[12] Kirişci M. A Case Study for medical decision making with the fuzzy soft sets. Afrika
Matematika 2020; 31: 557-564. doi:10.1007/s13370-019-00741-9

[13] Kirişci M. Ω− Soft Sets and medical decision-making application. International Journal
of Computer Mathematics, 2021;98(4):690–704.

[14] Kirişci M., Simşek N. Decision making method related to Pythagorean Fuzzy Soft Sets
with infectious diseases application. Journal of King Saud University - Computer and
Information Sciences 2021; 34(8), 5968–5978, doi:10.1016/[Link].2021.08.010

[15] Kirişci M., Demir I., Şimşek N. Fermatean Fuzzy ELECTRE Multi-Criteria Group
Decision-Making and Most Suitable Biomedical Material Selection. Artificial Intelli-
gence in Medicine 2022, 127, 102278, doi:10.1016/[Link].2022.102278

[16] Kirişci M. Correlation Coefficients of Fermatean Fuzzy Sets with a Medical Application.
J. Math. Sci. Model. 2022, 5(1), 16–23, doi: 10.33187/jmsm.1039613

[17] Şimşek N., Kirişci M. Incomplete Fermatean Fuzzy Preference Relations and Group
Decision Making. Topological Algebra and its Applications 2022.

[18] Kirişci M. Fermatean Hesitant Fuzzy Sets with Medical Decision Making Application.
doi:10.21203/[Link]-1151389/v2.

[19] Kirişci M., Şimşek N. Interval-Valued Fermatean Hesitant Fuzzy Sets and Infectious
Diseases Application. doi: 10.21203/[Link]-1273874/v1.

[20] Kirişci, M. New Entropy and Distance Measures for Fermatean Fuzzy Soft
Sets with Medical Decision-Making and Pattern Recognition Applications.
[Link]
30

[21] Kirişci M. Fermatean Fuzzy Type Variance, Covariance and Correlation Coefficients
with their Application [Link]
Fermatean_Fuzzy_Type_Variance_Covariance_and_Correlation_Coefficients_
with_their_Application
[22] Kirişci M. New Cosine Similarity and Distance Measures for Fermatean Fuzzy Sets and
TOPSIS Approach. Knowledge and Information Systems 2022.
[23] Liang Z., Shi P. Similarity measures on intuitionistic fuzzy sets. Pattern Recognition
Letters 2003; 24 (15): 2687-2693. [Link]/10.1016/S0167-8655(03)00111-9
[24] Liu S.T., Kao C. Fuzzy measures for correlation coefficient of fuzzy numbers. Fuzzy
Sets and Systems 2002; 128 (2): 267-275. [Link]/10.1016/S0165-0114(01)00199-3
[25] Liu D., Liu Y., Wang L. Distance measure for Fermatean fuzzy linguistic term sets
based on linguistic scale function: An illustration of the TODIM and TOPSIS methods.
International Journal of Intelligent Systems 2019; 34(11), 2807-2834
[26] Maji P. K., Biswas R., Roy A. R. Fuzzy soft sets. Journal of Fuzzy Mathematics, 2001;
9, 589–602.
[27] Maji P. K., Biswas R., Roy A. R. Intuitionistic fuzzy soft sets, Journal of Fuzzy Math-
ematics 2001; 9, 677–692.
[28] Molodtsov D. Soft set theory-First results. Computers & Mathematics with Applica-
tions 1999; 37(4–5), 19–31.
[29] OrthoInfo, Fracture After Total Hip Replacement, [Link]
treatment/hip-resurfacing/
[30] OrthoInfo, Fracture After Total Hip Replacement, [Link]
diseasesconditions/fracture-after-total-hip-replacement/
[31] Peng X., Yang Y., Song J. Pythagorean fuzzy soft set and its application. Computer
Engineering 2015; 41, 224–229.
[32] Peng X., Selvachandran G. Pythagorean fuzzy set: state of the art and future directions.
Artificial Intelligence Review 2019; 52 (3): 1873-1927. doi:10.1007/s10462-017-9596-9.
[33] Abdullah S., Qiyas M., Naeem M., Mamona, Liu Y. Pythagorean Cubic fuzzy
Hamacher aggregation operators and their application in green supply selection
problem. AIMS Mathematics 2022, 7(3): 4735–4766. doi: 10.3934/math.2022263

[34] Senapati T., Yager R.R. Fermatean fuzzy sets. Journal of Ambient Intelligence and
Humanized Computing 2020; 11 (2): 663-674.
[35] Senapati T., Yager R.R. Some new operations over Fermatean fuzzy numbers and ap-
plication of Fermatean fuzzy WPM in multiple criteria decision making. Informatica
2019; 30 (1): 391-412.
[36] Senapati T., Yager R.R. Fermatean fuzzy weighted averaging/geometric operators and
its application in multi-criteria decision-making methods. Engineering Applications of
Artificial Intelligence 2019; 85: 112-121. doi:10.1016/[Link].2019.05.012.
[37] Shahzadi G., Akram M. Group decision-making for the selection of an antivirus mask
under fermatean fuzzy soft information. Journal of Intelligent & Fuzzy Systems 2021;
40 (1): 1401-1416.
 31

[38] Akram M., Shahzadi G., Ahmadini A. A. H. Decision-Making Framework for an Effec-
tive Sanitizer to Reduce COVID-19 under Fermatean Fuzzy Environment. Journal of
Mathematics, 2020, Article ID 3263407, DOI:10.1155/2020/3263407
[39] Simşek N., Kirişci M. Incomplete Fermatean Fuzzy Preference Relations and
Group Decision Making. [Link]
Incomplete_Fermatean_Fuzzy_Preference_Relations_and_Group_Decision_
Making
[40] Smarandache F. Extension of soft set to hypersoft set, and then to plithogenic hypersoft
set. Neutrosophic Sets and Systems 2018; 22, 168–170.
[41] Vander W.D., Nachtegael M., Kerre E.E. A new similarity measure for image processing.
Journal of Computational Methods in Sciences and Engineering 2003; 3 (2): 209-222.
[42] Vander W.D., Nachtegael M., Kerre E.E. Using similarity measures and homogeneity
for the comparison of images. Image and Vision Computing 2004; 22 (9): 695-702.
[Link]/10.1016/[Link].2004.03.002
[43] Wei W.G., Wang H.J., Lin R. Application of correlation coefficient to interval-valued
intuitionistic fuzzy multiple attribute decision-making with incomplete weight informa-
tion. Knowledge and Information Systems 2011; 26: 337-349.
[44] Xu Z.S., Chen J., Wu J.J. Cluster algorithm for intuitionistic fuzzy sets. Information
Sciences 2008; 178 (19): 3775-3790. doi:10.1016/[Link].2008.06.008
[45] Yager R.R. Pythagorean fuzzy subsets. Proc Joint IFSA World Congress and NAFIPS
Annual Meeting,Edmonton, Canada, 2013.
[46] Yager R.R. Pythagorean membership grade in multicriteria decision
makng. IEEE Transactions on Fuzzy Systems 2014; 22 (4): 958-965.
doi:10.1109/TFUZZ.2013.2278989
[47] Yager R.R., Abbasov A.M. Pythagorean membership grades, complex numbers and
decision making. International Journal of Intelligent Systems 2013; 28 (5): 436-452.
[48] Yoon K., Hwang C.-L. Multiple Attribute Decision Making: Methods and Applica-
tions, SPRINGER-VERLAG BERLIN AN, 1981
[49] Yang Z., Garg H., Li X. Differential Calculus of Fermatean Fuzzy Functions: Continu-
ities, Derivatives, and Differentials. International Journal of Computational Intelligence
Systems, 2021,; 14(1), 282–294.
[50] Yahya M., Naeem M., Abdullah S., Qiyas M., Aamir M. A novel approach on
the intuitionistic fuzzy rough frank aggregation operator-based EDAS method
for multicriteria group decision-making. Complexity. 2021; Article ID 5534381,
[Link]

[51] Zadeh, L.A. Fuzzy sets. Information and Control 1965; 8(3), 338–353.
[52] Zhang H. M., Xu Z. S., Chen Q. On clustering approach to intuitionistic fuzzy sets.
Control Decision 2007; 22, 882–888.
[53] Zhang X.L., Xu Z.S. Extension of TOPSIS to multi-criteria decision making with
Pythagorean fuzzy sets. International Journal of Intelligent Systems 2014; 29(12): 1061–
1078.
 32

[54] Zulqarnain R. M., Xin X. L., Saeed M. Extension of TOPSIS method under intu-
itionistic fuzzy hypersoft environment based on correlation coefficient and aggregation
operators to solve decision making problem. AIMS Mathematics 2020; 6(3), 2732–2755.
[55] Zulqarnain R. M., Xin X. L., Saeed M. A Development of Pythagorean fuzzy hypersoft
set with basic operations and decision-making approach based on correlation coefficient.
Neutrosophic Sets and Systems, 2021; 40, 149–168.

View publication stats