FUZZY SOFT SETS AND ITS APPLICATION IN DIMENSION REDUCTION AND

MEDICAL DIAGNOSIS

A project submitted to the Bharathiar University in partial

fulfillments of the requirements for the award of the degree of

MASTER OF SCIENCE IN MATHEMATICS

Submitted by

MEHAVARSHINI.V.S

([Link]. 20PMAT013)

Under the Guidance and Supervision of

[Link] MSc.,[Link].,Ph.D.,SET.,PGDCA.,

DEPARTMENT OF MATHEMATICS

KONGU ARTS AND SCIENCE COLLEGE

DBT STAR COLLEGE SCHEME

An Autonomous Institution affiliated to Bharathiar University, Coimbatore

Re-accredited by NAAC

An ISO 9001:2015 Certified Institution

ERODE - 638 107.

MAY - 2022
CERTIFICATE
 CERTIFICATE

This is to certify that the project, entitled ”FUZZY SOFT SETS AND ITS APPLICATION IN

DIMENSION REDUCTION AND MEDICAL DIAGNOSIS” submitted to the Bharathiar

University in partial fulfillments of the requirements for the award of the degree of MASTER OF

SCIENCE IN MATHEMATICS is a record of work done by MEHAVARSHINI.V.S

([Link].20PMAT013), during the period 2020-2022 of her study in the Department of

Mathematics, Kongu Arts and Science College (Autonomous), Erode, under my supervision

and guidance and the project has not formed the basis for the award of any Degree / Diploma /

Associateship / fellowship or similar title to any candidate of any university.

Place : Erode Signature of the Guide

Date : ([Link])

Countersigned

HEAD OF DEPARTMENT PRINCIPAL

Date of Viva-voice Examination

INTERNAL EXAMINER EXTERNAL EXAMINER

DECLARATION
 DECLARATION

I, MEHAVARSHINI V.S, hereby declare that the project entitled ”FUZZY SOFT SETS AND

ITS APPLICATION IN DIMENSION REDUCTION AND MEDICAL DIAGNOSIS”

submitted to the Bharathiar University in partial fulfillments of the requirements for the award of the

degree of MASTER OF SCIENCE IN MATHEMATICS is a record of work done by me dur-

ing the period 2020-2022 of my study under the supervision and guidance of [Link]

MSc.,[Link].,Ph.D., SET.,PGDCA., Associate Professor, Department of Mathematics, Kongu

Arts and Science College (Autonomous), Erode, and that it has not formed the basis for the award

of any Degree/ Diploma/ Associateship /fellowship or similar title to any candidate of any university.

Signature of the Candidate

(MEHAVARSHINI.V.S)
ACKNOWLEDGEMENT
 ACKNOWLEDGEMENT

I thankfully acknowledge my indebtedness to Thiru [Link] B.E.,M.B.A.,M.S.,

Correspondent and [Link] [Link].,M.B.A.,[Link].,PGDCA.,Ph.D., Principal, Kongu Arts

and Science College (Autonomous), Erode , for providing me all facilities to complete my project suc-

cessfully in time.

My sincere and heartful thanks goes to my guide [Link] MSc.,

[Link].,Ph.D.,SET.,PGDCA.., Associate Professor ,Department of Mathematics, Kongu Arts and

Science College (Autonomous), Erode , who evinced keen interest and spent her time in providing guid-

ance that has resulted in the present form and gave me an opportunity to do this project.

I am grateful to thank [Link] [Link].,[Link].,Ph.D.,[Link].,

PGDCA., Associate Professor and Head, Department of Mathematics, Kongu Arts and Science College

(Autonomous), Erode , for his motivation and encouragement to complete this project successfully.

I extend my thanks to all the FACULTY MEMBERS of the Department of Mathematics,

Kongu Arts and Science College (Autonomous), Erode, for the timely help rendered by them during the

course of my study.

I express my heartful thanks to MY PARENTS AND FRIENDS for their encouragement,

kind help and cooperation.

[Link]
ABSTRACT
 ABSTRACT

In our daily life, we often come across various problems related to the high dimen-
sionality of data. In such type of problems, irrelevant and superfluous data along with
useful data is also present. Thus, dimensionality reduction has found wide applications in
data analysis and management. In recent years, the issue of dimensionality reduction in a
fuzzy situation has also gained importance and has invited attention of researchers. Var-
ious techniques and theories have been developed to solve these types of problems. Some
of these techniques are based on probabilistic approach and others are non-probabilistic
approach. For finding coherent and logical solution to various real life problems contain-
ing uncertainty, impreciseness and vagueness, fuzzy soft set theory is gaining significance.
In this project concept of fuzzy soft set has been defined as hybridization of fuzzy set
and soft set theory. A new technique is proposed to convert the soft set table into fuzzy
soft set table and has been applied in dimension reduction of big data. An application of
fuzzy soft set has also been studied in Medical Diagnosis done by sanchez’s approach.
CONTENTS
 CONTENTS

CHAPTER NO. TITLE PAGE NO.

INTRODUCTION 1

I PRELIMINARIES 4

STUDY OF OPERATIONS
II 12
IN FUZZY SOFT SETS

A NEW TECHNIQUE FOR

III 25
FINDING THRESHOLD VALUE

IV APPLICATIONS 28

CONCLUSION 37

BIBLIOGRAPHY 38
INTRODUCTION
 INTRODUCTION

Various real life problems in engineering, social and medical sciences,

economics etc. Engage imprecise and enormous data and their solution

concern the use of mathematical principles based on uncertainty and im-

[Link] deal with such systems in an effective way a number of theories

have been proposed in recent years. Some of these theories are probability,

game theory, fuzzy sets, intuitionistic fuzzy sets, etc.

The most suitable theory for dealing with uncertainty is the theory

of fuzzy sets developed by Zadeh [16] in 1965. Another one is rough set

theory pioneered by Pawlak in 1982, which is also a momentous tactic to

modeling vagueness. This theory has been successfully applied to many

fields such as machine learning, data mining, data analysis, medicine etc.

In 1999, Molodtsov [9] introduced a general mathematical tool known

as soft set theory to handle the objects which have been defined using a very

loose and hence very general set of characteristics, which was completely

a new approach for modeling vagueness and uncertainties.

In addition to defining the fundamental outlines of soft set theory,

Molodtsov [9] also illustrated how soft set theory is free from parameter-

ization insufficiency condition of fuzzy set theory, rough set theory, prob-

ability theory and game theory. Soft set theory is a universal framework,

1
as various traditional models emerge as unique case of soft sets theory.

Soft set theory has prospective for application in resolving realistic

problems in economics, engineering, environment, social science, medical

science and business management. Yukksel [15] have applied soft sets in

diagnosis of the prostate cancer risk. The absence of any restriction on

the approximate description in soft set theory makes this theory very con-

venient and easily applicable. Thus in recent years, research on soft set

theory has been dynamic and impressive movement has been achieved.

Later on, Maji[5-6] studied the theory of soft sets and also promote a

hybrid model known as fuzzy soft set [7], which is a combination of soft set

and fuzzy set. In [6] they also studied the intuitionistic fuzzy soft set. The

concept of fuzzy soft set introduced by Maji and Roy [5] was generalized

by Majumdar and Samanta .

Extreme data dimensionality reductions becomes the centre of cu-

riosity to a significant point of study in various fields of application (refer

to Gupta and Sharma). A number of techniques have been proposed by

the researchers and authors related to dimensionality reduction. Hooda [3]

studied dimension reduction in multivariate analysis by using maximum

entropy criterion.

2
 Chen[1] present a new definition of parameterization reduction in

soft sets and compare this definition to the associated concept of attributes

reduction in rough set theory. In this project we have studied about fuzzy

soft set based approach to reduce the dimensionality of data. The novel

method of dimensionality reduction involves constructions of binary infor-

mation table from soft sets and fuzzy soft sets in a parametric sense for

dimensionality reduction.

3
CHAPTER I
 CHAPTER I

PRELIMINARIES

Definition 1.1 SET

Set theory is the mathematical theory of well-determined collections,

called sets, of objects that are called members, or elements, of the set. Pure

set theory deals exclusively with sets, so the only sets under consideration

are those whose members are also sets.

Definition 1.2 CRISP SET

A set defined using a characteristic function that assigns a value

of either 0 or 1 to each element of the universe,it shows a full member-

ship [Link] is bi-valued function [Link] membership means true or

false,yes or no,0 or 1.

Figure 1: Example of Crisp set

Eg: she is 18 years old

Rahul is 1.6m tall

4
Definition 1.3 CONCEPT OF FUZZY

A fuzzy is concept of which the boundaries of application can vary

considerably according to context or conditions, instead of being fixed once

and for [Link] mathematics is branch of Mathematics including fuzzy

set theory and fuzzy logic the deals with partial inclusion of elements is a

set on a spectrum, a opposed to simple binary ”yes” or ”no”( 0 or 1)in-

[Link] are types of fuzzy called fuzzy logic,fuzzy sets, fuzzy graphs,

fuzzy topology, fuzzy field theory, fuzzy orderings, fuzzy geometry, etc.

Definition 1.4 FUZZY SET

The concept of fuzzy set is an extension of the concept of a crisp

set. Just as a crisp set on a universal set X is defined by its characteristic

function form X to {0 to 1},simply that fuzzy set define the value between

0 and 1 means in the interval[0,1]. It specifies the degree to which some-

thing is [Link] shows Partial membership. It is infinte valued function

logic. Partial membership means true to false, yes to no, 0 to 1.

Figure 2: Example of Fuzzy set

5
 Eg: she is about 18 years old.

Rahul is about 1.6m tall.

Definition 1.5 SOFT SET

Let X be an initial universe set and E be the set of parameters and

is subset of X. Then the pair (F,A)is called a soft set(over X),if F mapping

of A into the power set of X, i.e F:A→ p(X).Obviously,for given universe

X a soft set is parameterized family of subsets over X.

Example:

Let X = {m1 , m2 , m3 , m4 , m5 } be the set of mobiles under con-

sideration and E be the set of parameters,A=E={e1 =expensive, e2 =good

quality, e3 =cheap, e4 =stylish, e5 =latest}. The soft set (F,E) describes the

attractiveness of the mobile as

(F,E)={{Expensive=m − 1, m2 , m3 },{Good quality = m3 , m5 },

{cheap = m3 , m4 , m5 }, {stylish=m1 , m2 , m3 , m4 , m5 },{latest=m1 , m3 , m4 , m5 }}

X/E e1 e2 e3 e4 e5

m1 1 0 0 1 1

m2 1 0 0 1 0

m3 0 1 1 1 1

m4 0 0 1 1 1

m5 1 1 1 1 1

Table 1: Example of soft set

6
Definition 1.6 FUZZY SOFT SET

Fuzzy soft set is hybridization of fuzzy set and soft sets, by[7] de-

fined fuzzy soft sets. Actually concept of fuzzy soft set is extension of crisp

soft [Link] X be the universal set and F(x) be the set of all fuzzy subsets

of X, Let E be a set of parameters and A is a subset of E, a pair (F,A)is

called a fuzzy soft set, where F is a mapping from A to F(X).The a fuzzy

soft set(F,A) over X can be represented by the set of ordered pairs

(F,A)={(p, FA (p)) : p ∈ P, FA (p) ∈ F (X)}

Example:

we can take the same example as soft set for the fuzzy soft set then

we considered above m2 has stylish then it will not be possible to express it

with only the two numbers 0 and 1. In that case we can characterize it by

a membership function instead of the crisp number 0 and 1 the associated

with each element a real number in the interval[0,1].then fuzzy soft set can

be described as FE = {FE (e1) = {m1/0.36, m2/0.24, m3/0, m4/0, m5/0.6},

FE (e2) = {m1/0, m2/0, m3/0.32, m4/0, m5/0.40},

FE (e3) = {m1/0, m2/0, m3/0.48, m4/0.36, m5/0.6},

FE (e4) = {m1/0.6, m2/0.4, m3/0.8, m4/0.6, m5/1},

FE (e5) = {m1/0.48, m2/0, m3/0.64, m4/0.48, m5/0.8}

The Tabular representation of this fuzzy soft set FE is as shown below:

7
 X/E e1 e2 e3 e4 e5

m1 0.36 0.00 0.00 0.60 0.48

m2 0.24 0.00 0.00 0.40 0.00

m3 0.00 0.32 0.48 0.80 0.64

m4 0.00 0.00 0.36 0.60 0.48

m5 0.60 0.40 0.60 1.00 0.80

Table 2: Example of fuzzy soft set

Definition 1.7 FUZZY MATRIX

Fuzzy matrix is matrix whose entries are real numbers in the inter-

val [0,1], through out this we denote the unit interval .we sayX ∈[0,1],if

0 ≤ x ≤ 1. We also denote the unit interval as the Fuzzy interval.

Definition 1.8 FUZZY RELATION MATRIX

If a fuzzy relation R is a given in the form of fuzzy matrix, its el-

ements represent the membership values of this relation. That is, if the

matrix is denoted by MR, and the membership values by R(ij).

8
 In following Definitions, we study about some basic definitions of

Fuzzy soft sets.

Definition 1.9 FUZZY SOFT CLASS

Let U be a universe and E a set of attributes. Then the pair(U,E)

denotes the collection of all fuzzy soft sets on U with attributes from E

and is called fuzzy soft sets

Definition 1.10 FUZZY SOFT NULL SET

A soft set(F,A) over U is said to be null fuzzy soft set denoted by ϕ

if ∀ε ∈A,F(ε) is the null fuzzy set 0̄ of U where 0̄(x)=0∀x ∈ U.

Definition 1.11 FUZZY SOFT ABSOLUTE SET

A soft set(F,A) over U is said to be absolute fuzzy soft set denoted

e if ∀ε ∈A,F(ε) is the absolute fuzzy set 1̄ of U where 1̄ ∀(x)=1∀ xεU
by A

Definition 1.12 FUZZY SOFT SUBSET

For two fuzzy soft sets(F,A) and (G,B) in a fuzzy soft class (U,E),

we say that (F,A) is a fuzzy soft subset of(G,B), if

i)A⊆ B

ii)For all ε ∈A,F (ε) ⊆ G(ε) and its written as (F,A)⊆(G,B)

9
Definition 1.13 UNION OF FUZZY SOFT SETS

Union of two fuzzy soft sets(F,A) and (G,B) in a soft class (U,E)is

a fuzzy sot set (H,C) where C=A∪B ∀ε ∈C


H(ε)= F(ε), if ε ∈ A − B

G(ε), if ε ∈ B − A

F(ε) ∪ G(ε), if ε ∈ A∩B

Thus ,(F, A)∪

e (G, B)=(H,C)

Definition 1.14 INTERSECTION OF FUZZY SOFT SETS

Intersection of two fuzzy soft sets(F,A) and (G,B) in a soft class

(U,E) is a fuzzy soft set(H,C) where C=A∩B and ∀ε ∈C, H(ε)=F(ε) or

G(ε)(as both are same fuzzy set) and in written as ,(F, A)∩
e (G, B)=(H,C)

Definition 1.15 INTERSECTION OF FUZZY SOFT SETS RE-

DEFINED

Let (F,A) and (G,B) be two fuzzy soft sets in a soft class(U,E) with

A∩B6= φ.Then Intersection of fuzzy soft sets (F,A) and (G,B) in a soft

class (U,E) is a fuzzy soft set (H,C) where C=A∩B and∀ε ∈C,

H(ε)=F(ε) ∩ G(ε).We write (F,A)∩

e (G,B)=(H,C).

10
Definition 1.16 COMPLEMENT OF FUZZY SOFT SET

The complement of the fuzzy soft set (F,A) is denoted by (F, A)c

and is defined by (F, A)c = (F c , A) where F c : A → P̃ (U) is a mapping

given by F c (α)=[F (α)]c , ∀α ∈A

Definition 1.17 AND OPERATION OF FUZZY SOFT SET

If (F,A) and (G,B) be the two fuzzy soft sets, then ”(F,A)AND(G,B)”

is a fuzzy soft set denoted by (F,A)∧(G,B) and is defined by

(F,A)∧(G,B)=(H,A× B,where H(α, β)=F(α)∩G(β), ∀α ∈A and∀β ∈B,

where ∩ is the operation intersection of two fuzzy soft sets.

Definition 1.18 OR OPERATION OF FUZZY SOFT SET

If (F,A) and (G,B) be the two fuzzy soft sets, then ”(F,A)OR(G,B)”

is a fuzzy soft set denoted by (F,A)∨(G,B) and is defined by

(F,A)∨(G,B)=(K,A× B,where K(α, β)=F(α)∪G(β), ∀α ∈A and∀β ∈B,

where ∪ is the operation intersection of two fuzzy soft sets.

11
CHAPTER I
 CHAPTER II

A STUDY ON THE OPERATIONS IN FUZZY SOFT SETS

In this chapter, we shall endeavour to study the basic operations and re-

sults available in the literature of fuzzy soft sets. Ahmad and Kharal has

pointed out that the null (resp., absolute) fuzzy soft set as defined by Maji

et al., is not unique in a fuzzy soft class (U,E), rather it depends upon

A ⊆ E. As such whenever we would refer to a fuzzy soft null set or fuzzy

soft absolute set, we would refer to the set A ⊆ E of parameters under

consideration. In case of soft sets, this has already been pointed out by Ge

and Yang in and accordingly they have studied some basic results regarding

soft sets. Here, we put forward some results of fuzzy soft sets in our way.

We are using the notation (φ, A) to represent the fuzzy soft null set with

respect to the set of parameters A and the notation (U,A) to represent the

fuzzy soft absolute set with respect to the set of parameters A.

12
Proposition 1:

(φ, A)c = (U, A)

Proof :

Let (φ, A) = (F, A)

Then ∀ε ∈ A,

F (ε) = {(x, µF (ε) (x)) : x ∈ U }

= {(x, 0) : x ∈ U }

(φ, A)c = (F, A)c = (F c , A), where

∀ε ∈ A, F c (ε) = (F (ε))c

= {(x, µF (ε) (x)) : x ∈ U }c

= {(x, 1 − µF (ε) (x)) : x ∈ U }

= {(x, 1 − 0) : x ∈ U }

= {(x, 1) : x ∈ U }

=U

Thus, (φ, A)c = (U, A)

Proposition 2:

(U, A)c = (φ, A)

Proof :

13
Let (U,A) = (F,A)

Then ∀ε ∈ A,

F (ε) = {(x, µF (ε) (x)) : x ∈ U }

= {(x, 1) : x ∈ U }

(U, A)c = (F, A)c = (F c , A), where

∀ε ∈ A, F c (ε) = (F (ε))c

= {(x, µF (ε) (x)) : x ∈ U }c

= {(x, 1 − µF (ε) (x)) : x ∈ U }

= {(x, 1 − 1) : x ∈ U }

= {(x, 0) : x ∈ U }

=φ
Thus, (U, A)c = (φ, A)

Proposition 3:

˜ (φ, A) = (F, A)
(F, A)∪

Proof:

We have

(F, A) = {(ε, (x, µF (ε) (x))) : x ∈ U }∀ε ∈ A

(φ, A) = {(ε, (x, 0)) : x ∈ U }∀ε ∈ A

14
 ˜ (φ, A) = {(ε, (x, max(µF (ε) (x), 0))) : x ∈ U }∀ε ∈ A
(F, A)∪

= {(ε(x, µF (ε) (x))) : x ∈ U }∀ε ∈ A

= (F, A)
˜ (φ, A) = (F, A)
Thus, (F, A)∪

Proposition 4:

˜ (U, A) = (U, A)
(F, A)∪

Proof:

We have

(F, A) = {(ε, (x, µF (ε) (x))) : x ∈ U }∀ε ∈ A

(U, A) = {(ε, (x, 1)) : x ∈ U }∀ε ∈ A

˜ (U, A) = {(ε, (x, max(µF (ε) (x), 1))) : x ∈ U }∀ε ∈ A

(F, A)∪

= {(ε, (x, 1)) : x ∈ U }∀ε ∈ A

= (U, A)
˜ (U, A) = (U, A)
Thus (F, A)∪

15
Proposition 5:

˜ (φ, A) = (φ, A)
(F, A)∩

Proof:

We have

(F, A) = {(ε, (x, µF (ε) (x))) : x ∈ U }∀ε ∈ A

(φ, A) = {(ε, (x, 0)) : x ∈ U }∀ε ∈ A

˜ (φ, A) = {(ε, (x, min(µF (ε) (x), 0))) : x ∈ U }∀ε ∈ A

(F, A)∩

= {(ε, (x, 0)) : x ∈ U }∀ε ∈ A

= (φ, A)

˜ (φ, A) = (F, A)
Thus (F, A)∩

Proposition 6:

˜ (U, A) = (F, A)
(F, A)∩

Proof:

We have

16
 (F, A) = {(ε, (x, µF (ε) (x))) : x ∈ U }∀ε ∈ A

(U, A) = {(ε, (x, 1)) : x ∈ U }∀ε ∈ A

˜ (U, A) = {(ε, (x, min(µF (ε) (x), 1))) : x ∈ U }∀ε ∈ A

(F, A)∪

= {(ε, (x, µF (ε) (x))) : x ∈ U }∀ε ∈ A

= (F, A)
˜ (U, A) = (U, A)
Thus, (F, A)∪

Proposition 7:

˜ (φ, B) = (F, A) if and only if B ⊆ A

(F, A)∪

Proof :

We have for (F,A),

F (ε) = {(x,µF (ε) (x)):x ∈ U } ∀ε ∈ A

Also, let (φ, B) = (G, B), Then

G(ε) = {(x, 0) : x ∈ U }∀ε ∈ B

˜ (φ, B) = (F, A)∪

Let (F, A)∪ ˜ (G, B) = (H, C), where C = A ∪ B and

∀ε ∈ C,

17
 

{(x, µF (ε) (x)) : x ∈ U }, if ε ∈ A − B







H(ε) = {(x, µG(ε) (x)) : x ∈ U }, if ε ∈ B − A





{(x, max(µF (ε) (x), µG(ε) (x))) : x ∈ U }, if ε ∈ A ∩ B





{(x, µF (ε) (x)) : x ∈ U }, if ε ∈ A − B







= {(x, 0) : x ∈ U }, if ε ∈ B − A





{(x, max(µF (ε) (x), 0)) : x ∈ U }, if ε ∈ A ∩ B





{(x, µF (ε) (x)) : x ∈ U }, if ε ∈ A − B







= {(x, 0) : x ∈ U }, if ε ∈ B − A





{(x, (µF (ε) (x)) : x ∈ U }, if ε ∈ A ∩ B



Let B ⊆ A Then



{(x, µ

F (ε) (x)) : x ∈ U }, if ε ∈ A − B
H(ε) =

{(x, µF (ε) (x)) : x ∈ U }, if ε ∈ A ∩ B


= F (ε)∀ε ∈ A

˜ (φ, B) = (F, A)
Conversely, let (F, A)∪

Then, A=A ∪ B ⇒ B ⊆ A

18
Proposition 8:

˜ (U, B) = (U, B) if and only if A ⊆ B

(F, A)∪

Proof :

We have for (F,A)

F (ε) = {(x, µF (ε) (x)) : x ∈ U }, if ε ∈ A

Also, let (U,B)=(G,B), Then

G(ε) = {(x, 1) : x ∈ U }, ∀ ε ∈ B
˜ (U, B) = (F, A)∪
Let (F, A)∪ ˜ (G, B) = (H, C), where C = A ∪ B and

∀ε ∈ C,



{(x, µF (ε) (x)) : x ∈ U }, if ε ∈ A − B







H(ε) = {(x, µG(ε) (x)) : x ∈ U }, if ε ∈ B − A





{(x, max(µF (ε) (x), µG(ε) (x))) : x ∈ U }, if ε ∈ A ∩ B





{(x, µF (ε) (x)) : x ∈ U }, if ε ∈ A − B







= {(x, 1) : x ∈ U }, if ε ∈ B − A





{(x, max(µF (ε) (x), 1)) : x ∈ U }, if ε ∈ A ∩ B



19
 

{(x, µF (ε) (x)) : x ∈ U }, if ε ∈ A − B







= {(x, 1) : x ∈ U }, if ε ∈ B − A





{(x, 1) : x ∈ U }, if ε ∈ A ∩ B


Let A ⊆ B

Then 

{(x, 1) : x ∈ U }, if ε ∈ B − A

H(ε) =

{(x, 1)) : x ∈ U }, if ε ∈ A ∩ B


= G(ε)∀ε ∈ B

˜ (U, B) = (U, B)
Conversely, let (F, A)∪

Then B=A ∪ B ⇒ A ⊆ B

Proposition 9:

˜ (φ, B) = (φ, A ∩ B)
(F, A)∩

Proof :

We have for (F,A)

F (ε) = {(x,µF (ε) (x)):x ∈ U } ∀ε ∈ A

20
Also, let (φ, B) = (G, B), Then

G(ε) = {(x, 0) : x ∈ U }∀ε ∈ B

˜ (φ, B) = (F, A)∩

Let (F, A)∩ ˜ (G, B) = (H, C), where C = A ∩ B and

∀ε ∈ C,

H(ε) = {(x, min(µF (ε) µG(ε) (x)) : x ∈ U }

= {(x, min(µF (ε) , 0)) : x ∈ U }

= {(x, 0) : x ∈ U }
˜ (φ, B) = (φ, A ∩ B)
Thus (F, A)∪

Proposition 10:

˜ (U, B) = (F, A ∩ B)
(F, A)∩

Proof:

We have for (F,A)

F (ε) = {(x, µF (ε) (x)) : x ∈ U }∀ε ∈ A

Also, let (U, B) = (G, B), Then

G(ε) = {(x, 1) : x ∈ U }∀ε ∈ B

˜ (φ, B) = (F, A)∩
Let (F, A)∩ ˜ (G, B) = (H, C),

where C = A ∩ B and ∀ε ∈ C,

21
 H(ε) = {(x, min(µF (ε) (x), µG(ε) (x))) : x ∈ U }

= {(x, min(µF (ε) (x), 1)) : x ∈ U }

= {(x, µF (ε) (x)) : x ∈ U }

˜ (U, B) = (F, A ∩ B)
Thus, (F, A)∩

Proposition (De Morgan Laws)

For fuzzy soft sets (F,A) and (G,A) over the same universe U, we have the

following -
˜ (G, A))c = (F, A)c ∩
1. ((F, A) ∪ ˜ (G, A)c

˜ (G, A)c = (F, A)c ∪

2.((F, A) ∩ ˜ (G, A)c

Proof:
˜ (G, A) = (H, A), where ∀ε ∈ A,
[Link] (F, A) ∪

H(ε) = F (ε) ∪ G(ε)

= {x, max(µF (ε) , µG(ε) )}

Thus

˜ (G, A))c = (H, A)c = (H c , A) where ∀ε ∈ A,

((F, A)∪

H c (ε) = (H(ε)c )

22
 = (F (ε) ∪ G(ε))c

= {x, 1 − max(µF (ε) , µG(ε) )}

= {x, min(1 − µF (ε) (x), 1 − µG(ε) (x))}

Again, (F, A)c ∩

˜ (G, A)c = (F c , A) ∩
˜ (Gc , A) = (I, A), say

Where ∀ε ∈ A,

I(ε) = F c (ε) ∩ Gc (ε)

= {x, min(µF c (ε) (x), µG(ε) (x))}

= {x, min(1 − µF (ε) (x), 1 − µG(ε) (x))}

˜ (G, A))c = (F, A)c ∩

Thus ((F, A)∪ ˜ (G, A)c

˜ (G, A) = (H, A), where ∀ε ∈ A,

[Link] (F, A)∩

H(ε) = F (ε) ∩ G(ε)

= {x, min(µF (ε) , µG(ε) )}

Thus

˜ (G, A))c = (H, A)c = (H c , A), where ∀ ε ∈ A,

((F, A) ∩

H c (ε) = (H(ε))c

= (F (ε) ∩ G(ε))c

23
 = {x, 1 − min(µF (ε) , µG(ε) )}

= {x, max(1 − µF (ε) (x), 1 − µG(ε) (x))}

Again, (F, A)c ∪

˜ (G, A)c = (F c , A) ∪
˜ (Gc , A) = (I, A), say

Where ∀ε ∈ A,

I(ε) = F c (ε) ∪ Gc (ε)

= {x, max(µF c (ε) (x), µGc (ε) (x))}

= {x, max(1 − µF c (ε) (x), 1 − µGc (ε) (x))}

˜ (G, A))c = (F, A)c ∩

Thus ((F, A)∪ ˜ (G, A)c

24
CHAPTER III
 CHAPTER III

A NEW TECHNIQUE FOR FINDING THRESHOLD VALUE

In this chapter we introduce a new concept of dimensionality reduction

by using the fuzzy soft sets approach via soft sets. Here we also propose

a process by which we convert binary valued information table of soft

sets into the object oriented membership valued information table of fuzzy

soft sets. By this process, we assign the membership value of objects

with respect to parameters in fuzzy soft sets by using soft set information.

Before discussing our proposed method we first define some terms.

Given a soft set (Ω , E)with tabular presentation,

X={m1 , m2 , m3 ....mn }is the object set, E = e1 , e2 , . . . , em is the parameter

set, and mij are the entries in the table of (Ω, E).

Definition 3.1

For soft set (Ω, E), X = (m1 , m2 . . . , mn } , E = {e1 , e2 . . . , em } we denote

P
TE (mi ) = j mj as an oriented- object sum and Ri = (TE (mi ) /|E|) is

oriented-object grade with respect to parameters.

25
Definition 3.2

For soft set (Ω, E), X = (m1 , m2 , . . . , mn } , E = {e1 , e2 , . . . , en } we denote

P
SX (ej ) = j mj as an oriented- parameter sum and Cj = (SX (ej ) /(|X|)

is oriented-parameter grade with respect to objects.

Definition 3.3

For soft set (Ω, E)X = {m1 , m2 . . . , mn } , E = {e1 , e2 , . . . , em } we construct

a fuzzy soft set FE over X by assigning membership value of objects with re-

spect to parameter by using definition 3.1 and Definition 3.2. Let vij be the

membership value of objects in fuzzy soft set FE then vj = (mij × Ri × Cj )

such that vij ∈ [0, 1].

Definition 3.4

For Fuzzy soft set FE , X= {m1 , m2 , . . . , mn }, E = {e1 , e2 , . . . , em } we do-

P
nate Oi = ( j vij )/|E| oriented object grade with respect to parameters

Definition 3.5

For Fuzzy soft set FE ,X={m1 , m2 , . . . , mn }, E = {e1 , e2 , . . . , em } we denote

P
Ej = ( i vij )/|X| oriented parameter grade with respect to objects.

For Fuzzy soft set represent in example 3 the values of

Oj = 0.29, 0.13, 0.45, 0.29, 0.68 for i=1 to 5 respectively and the value of

Ej = 0.24, 0.14, 0.29, 0.68, 0.48 for j=1 to 5 respectively.

26
In proposed technique we define a term known as threshold value.

Definition 3.6

P P
For Fuzzy soft set FE we denote T = ( i j vij /(|X| × |E|) as threshold

value of fuzzy soft set. For fuzzy soft set represent in example 2.4, the

threshold value defined in definition 3.6 is 0.37.

3.1 Proposed Algorithm

The step wise process of proposed technique is as follows:

ˆ Input the soft set (Ω, E)

ˆ Construct the binary valued information table of soft set(ΩE).

ˆ Convert soft set (Ω , E)into fuzzy soft set FE and construct fuzzy soft

set table by using soft set table by proposed technique.

ˆ Find the value of Grade membership Oi of ith object and Ej of jth

parameter in fuzzy soft set.

ˆ Find the threshold value ’ T ’ of fuzzy soft sets

ˆ Remove those rows for which Oi < T and column for which Ej > T in

fuzzy soft set table.

ˆ The new table is our desired dimensionality reduced table.

27
CHAPTER IV
 CHAPTER IV

APPLICATIONS

In this chapter we discuss about two different applications of fuzzy soft set

in dimensionality reduction to clear a fuzzy situation of selecting a suitable

car from different available features and diagnosing a disease from various

symptoms of two diseases.

4.1 APPLICATION OF NEW TECHNIQUE IN DIMENSION-

ALITY REDUCTION

Here we analyze again the example presented by Maji in [7-8] and is dis-

cussed by Chen [16]. Let U = {h1 , h2 , h3 , h4 , h5 , h6 } be a set of six houses,

E ={ expensive, beautiful, wooden, cheap, in green surroundings, modern,

in good repair, in bad repair} be the set of parameters.

Let Mr. X is interested to buy a house on the following parameters subset

P = {beautiful, wooden, cheap, in green surroundings, in good repair}.

Let (e1 , e2 , e3 , e4 , e5 ) be symbolic representation of the set P . Boolean-

valued information system table gives the soft set as in Table 3.

28
 U/P p1 p2 p3 p4 p5

h1 1 1 1 1 1

h2 1 1 1 1 0

h3 1 0 1 1 1

h4 1 0 1 1 0

h5 1 0 1 0 0

h6 1 1 1 1 1

Table 3: Boolean-valued soft set

Next we determine oriented object grade Ri and oriented parameter

grade Ci of given soft set by using the table 2 and applying the proposed

technique,then table 3 transformed into the following Table 4.

U/P p1 p2 p3 p4 p5 Ri

h1 1 1 1 1 1 1

h2 1 1 1 1 0 4/5

h3 1 0 1 1 1 4/5

h4 1 0 1 1 0 3/5

h5 1 0 1 0 0 2/5

h6 1 1 1 1 1 1

Cj 1 1/2 1 5/6 1/2

Table 4: Ri and Cj grades of soft set

Transform Table 4 into the fuzzy soft set table by assigning membership

value vij of object using the proposed method where vij = (mij × Ri × Cj )

such that vij ∈ [0, 1], thus the transformed table is as:

29
 U/P p1 p2 p3 p4 p5

h1 1.00 0.50 1.00 0.83 0.50

h2 0.80 0.40 0.80 0.66 0

h3 0.80 0 0.80 0.66 0.40

h4 0.60 0 0.60 0.50 0

h5 0.40 0 0.40 0 0

h6 1.00 0.50 1.00 0.83 0.50

Table 5: Example of fuzzy soft set

P
Now determine oriented-object grade Hi = ( j vij )/|P | and oriented-
P
parameter grade Pj = ( i vij )/|U | of fuzzy soft set table where i = 1to|U|

and j = 1to|P|. Then by applying this process, Table 5 is transformed into

new table as Table 6.

P P
Now determine T = ( i j vij )/|U | × |P |) threshold value of fuzzy

soft sets, thus by calculation we get value of T = 0.52. Now remove those

rows for which Hi < T and those columns for which Pj > T. Now after

removing

30
 U/P p1 p2 p3 p4 p5 Hi

h1 1.00 0.50 1.00 0.83 0.50 0.77

h2 0.80 0.40 0.80 0.66 0 0.53

h3 0.80 0 0.80 0.66 0.40 0.53

h4 0.60 0 0.60 0.50 0 0.34

M5 0.40 0 0.40 0 0 0.16

M6 1.00 0.50 1.00 0.83 0.50 0.77

Pj 0.75 0.23 0.75 0.58 0.23

Table 6: Hi andPj grades of fuzzy soft set

The corresponding rows and columns Table 4 transform into new

Table 5 and that is desired reduced table:

U/P p2 p5 Hi

h1 0.50 0.50 0.77

h2 0.40 0 0.53

h3 0 0.40 0.53

h6 0.50 0.50 0.77

P 023 023

Table 7: Data Size Reduced Information

As can be seen above, data size has been reduced to approximately

75%. Yet after the process of reduction the reduced dataset maintain the

same decision partition stated by Maji [5], that Mr. X prefers the house

h1 and h6 .

31
 4.2 APPLICATION IN MEDICAL DIAGNOSIS

Here we study an application of fuzzy soft sets in medical diagnosis

following Sanchez’s approach [13].

Methodology

Suppose S is a system of symptoms, D is a set of diseases and P is a set

of patients. Then A fuzzy soft set (F, D) is constructed over S, where

F is a mapping from D to lS . This fuzzy soft set gives a relation matrix

R1 (say) and is called symptom-disease matrix. Its complement (F, D)c

gives another relation matrix R2 ( say) and is called non symptom disease

matrix. These matrices are referred to ’Soft Medical Knowledge’.

Again we construct another Fuzzy Soft Set (F1 ,S) over P , where

F1 is a mapping from S to P. This fuzzy set gives a relation matrix Q

called patient-system matrix. Then we obtain two new relation matrices

are given by T1 = Qo1 and T2 = QoR2 and are called symptom-patient

matrix and non symptom-patient matrix respectively. These matrices are

obtained by combining the relation matrices R1 and R2 with Q and follow-

ing max[min]rule and their membership values are given by

µR1 (pi , dj ) = ∪[µq , (pi , dj ) ∩ µR1 (ek , dj ] and

µR2 (pi , dj ) = ∪[µq , (pi , dj ) ∩ µR2 (ek , dj ], where ∪=max and ∩=min. In case

32
max {µR1 (pi , dj ) − µR2 (pi , dj )} occurs for exactly (pi , dk ), only then we con-

clude that the acceptable diagnostic hypothesis for patient pi is the disease

dk , however, if there is a tie, the process is to be repeated for patient pi by

reassessing the symptoms.

A Case Study

Suppose there are three patient’s p1 , p2 and p3 in a hospital with symptoms

temperature, headache, cough and stomach problem. Let the possible dis-

eases relating to these symptoms be viral fever and malaria. We consider

the setS = {e1 , e2 , e3 , e4 }, where e1 , e2 , e3 , e4 represent symptoms tempera-

ture, headache, cough and stomach problem respectively and D = {d1 , d2 },

where d1 and d2 represent the parameters viral fever and malaria respec-

tively.

Next, we suppose that F (d1 ) = {e1 /.8, e2 /.5, e3 /.4, e4 /.3} ,

F (d2 ) = {e1 /.6, e2 /.4, e3 /.3, e4 /.9}. Hence the fuzzy soft set (F, D) is a pa-

rameterized family of all fuzzy sets over the set S given by {F (d1 ) , F (d2 )}

and determined from expert medical documentation. Thus the fuzzy soft

set (F,D) gives an approximate description of the soft medical knowledge

of the two diseases and symptoms and are represented by the following two

relation matrices R1 and R2

33
 d1 d2
 
e1  .8 .6 
 
e2  .5 .4 
 
R1 = 



e3 
 .4 .3 

 
e4 .3 .9

d1 d2
 
e1  .2 .4 
 
e2  .5 .6 
 
R2 = 



e3 
 .6 .7 

 
e4 .7 .1

Let P={p1 , p2 , p3 }be the universal set where p1 , p2 andp3 represent the

patients under consideration and S = {e1 , e2 , e3 , e4 }, where e1 , e2 , e3 , e4

represent symptoms temperature,headache,cough and stomach problem

respectively be the parameters, Further suppose that

F (e1 )={p1 /.7, p2 /.8, p3 /.3},

F (e2 )={p1 /.5, p2 /.4, p3 /.5},

F (e3 )={p1 /.6, p2 /.5, p3 /.4},

F (e4 )={p1 /.4, p2 /.7, p3 /.5}.

Then fuzzy soft set(F,S) represents a relation matrix Q called patient-

symptoms matrix and is given by

34
 e1 e2 e3 e4
 
p1  .7 .5 .6 .4 
 
Q = p2  .8 .4 .5 .7 
 
 
 
p3 .3 .5 .4 .5

On combining the relation matrices R1 and R2 with Q and following

max[min] rule, we got two matrices T1 and T2 called patient disease and

patient-non disease matrices respectively and given by

d1 d2
 
p1  .7 .6 
 
T 1 = Q0 R1 = p2  .8 .7 
 
 
 
p3 .5 .5

d1 d2
 
p1  .6 .6 
 
T 1 = Q0 R1 = p2  .7 .5 
 
 
 
p3 .5 .5

d1 d2
 
p1  .1 0 
 
T1 − T2 = p2  .1 .2 
 
 
 
p3 0 0

35
Thus form matrixT1 − T2 we can infer that p1 are suffering from d1 and p2

is suffering from d2 and p3 is neither suffering from d1 nor d2 .

36
CONCLUSION
 CONCLUSION

In this project, a new technique of dimensionality reduction are discussed

by using fuzzy soft set theory via soft set theory. By using proposed tech-

nique in considered example taken by Maji [5] we see that data size has

been reduced to approx 75% and reduced data still maintain the same de-

cision partition stated by Maji [6]. The example illustrates our method

and shows that the new algorithm efficiently captures the reduction. We

have also studied an application of fuzzy soft set in medical diagnosis and

illustrated by a case study.

37
BIBLIOGRAPHY
 BIBLIOGRAPHY

[1] [Link] and [Link],”Onfuzzy Soft Sets”,Advances in Fuzzy Sys-

tem, Volume 2009.

[2] Chen D, Tsang ECC, Yeungand DS, WangX. The parameterization re-

duction of soft sets and its application Computer and Mathematics with

Applied. 2005;49:757-763

[3] Hooda DS, Hooda BK. Dimension reduction in multivariate analysis

using maximum entropy criterion. Journal of Statistics and Management.

2006;9(1): 175-183.

[4] Hooda DS, Vivek Raich. Fuzzy set theory and fuzzy control. Narosh

Publishers, New Delhi; 2016.

[5] Maji PK, Biswas R, Roy AR. Fuzzy soft sets. The Journal of Fuzzy

Mathematics. 2001;9(3):589-602.

[6] Maji PK, Biswas R, Roy AR. Intuitionistic fuzzy soft sets. The Journal

of Fuzzy Mathematics. 2001;9(3):677-692.

[7] Maji PK, Biswas R, Roy AR. An application of soft sets in a de-

cision making problem. Computer and Mathematics with Application.

2002;44:1077-1083.

[8] Maji PK, Roy AR. Soft set theory. Computers and Mathematics with

Applications. 2003;45:555-562.

[9] Molodstov DA. Soft set theory-first result. Computers and Mathematics

38
with Applications. 1999;37:19-31.

[10] Neogand TJ, Sut DK. A new approach to the theory of soft sets. In-

ternational Journal of Computer Applications. October2011; V ol32N o(2) :

pp1 − 6.

[11] [Link], [Link],”On Fuzzy Soft Complement and Related Proper-

ties”.International Journal of Energy, Information and communications.

[12] [Link], [Link],”On Union and Intersection of Fuzzy soft sets”,International

Journal of [Link],Vol2(5), 1160-1176

. [13] Sanchez E. Inverses of fuzzy relation and applications to possibility

distributions and medical diagnosis. Fuzzy Sets and Systems. 1979; 2(1) :

75 − 86.

[14] Xun Ge and Songlin Yang,”Investigations of some operations of soft

sets. World academy of of science Engineering and Technology ,75,2011,pp,1113-

1116.

[15] Yuksel S, Dizman T, Yildizdan G, Sert U. Application of soft sets to

diagnose the prostate cancer risk. J. Inequal. [Link]: 10.1186/1029-

241X-(2013)-229.

[16] Zadeh LA. Fuzzy sets. Information and Control. 1965;8:338-353.

39