STUDIES ON INTUTIONISTIC FUZZY INFORMATION MEASURE Ist International Conference of ISITA on Mathematical Modeling, Optimization and Information Technology SSCET, Badhani, Pathankot, Punjab 1 Surender Singh Assistant Prof. School of Mathematics, Faculty of Sciences Shri Mata Vaishno Devi University Katra –182320 (J & K) Email:[email protected]16 th -19 th Jan., 2015
28
Embed
STUDIES ON INTUTIONISTIC FUZZY INFORMATION MEASURE
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
STUDIES ON INTUTIONISTIC FUZZY INFORMATION MEASURE
Ist International Conference of ISITA on
Mathematical Modeling, Optimization and Information TechnologySSCET, Badhani, Pathankot, Punjab
Fuzziness, a feature of imperfect information, results from the lack of crisp distinction between the elements belonging and not belonging to a set (i.e. the boundaries of the set under consideration are not sharply defined).
The concept of fuzziness initiated by [Zadeh, 1965].
[De Luca and Termini, 1971] introduced some requirements which capture our intuitive comprehension of the degree of fuzziness in a fuzzy capture our intuitive comprehension of the degree of fuzziness in a fuzzy set and introduced concept of fuzzy entropy.
The term ‘Fuzzy entropy’ have been adopted due to an intrinsic similarity of equation to the one in the [Shannon, 1948] entropy.
Two functions measure fundamentally different types of uncertainty. Basically, the Shannon entropy measures the average uncertainty in bits associated with the prediction of outcomes in a random experiment.
3
The idea of intutionistic fuzziness initiated by [Atanassov, 1986] and incidentally affected all fields of study where ever concept of fuzziness was used. [Vlachos et al., 2007] derived an extension of De Luca- Termini’s entropy for IFSs. In this paper, we study measure of fuzziness for intuitionistic fuzzy sets in light of restricted equivalence functions and T-norm operator.
4
Definition 1.1.1[Bustince et al., 2006] Let cc ],1,0[]1,0[: is a fuzzy
negation iff:N1: 1)0( c and ,0)1( c
N2: yxifycxc ,)()( (monotonicity).
A fuzzy negation is strict, iff,N3: )(xc is continuous.
N4: yxifycxc ,)()( for all ].1,0[, yx
5
N4: yxifycxc ,)()( for all ].1,0[, yx
A strict fuzzy negation is involutive, iff,N5: ),())(( xcxcc ].1,0[x
The strict fuzzy negations that are involutive are called strong negations.
Definition 1.1.2[Bustince et al., 2006] A function ]1,0[]1,0[: 2 REF is called a
restricted equivalence function, if it satisfies the following conditions:R1: ),(),( xyREFyxREF for all ];1,0[, yx
R2: 1),( yxREF iff ;yx
R3: 0),( yxREF iff 1x and 0y or 0x and ;1y
R4: ))(),((),( ycxcREFyxREF for all ],1,0[, yx c being a strong negation;
R5: For all ],1,0[,, zyx if ,zyx then ),(),( zxREFyxREF
and
6
and .),(),( zxREFzyREF
It can be proved that R5 is equivalent to: for all x, y, z, t [0, 1], if ,tzyx then .),(),( txREFzyREF
1.2 Intutionistic fuzzy sets
Let X be an ordinary non-empty, finite set. Then an IFS over X is
characterized by two mappings ]1,0[: Xg and ].1,0[: Xh
For each )(, xgXx can be interpreted as the degree to which x enjoys some
property P. Alternately, )(xh is the degree to which x does not enjoy the property P. Here, g and h are the generalizations of characteristic function of conventional set theory. There is nothing intutionally fuzzy about set X. Rather, the fuzziness lies in the degree of compatibility and the degree of
7
Rather, the fuzziness lies in the degree of compatibility and the degree of incompatibility of the element Xx with property P. Definition 1.2.1. An intutionistic fuzzy set A defined on a universe X is given by [Atanassov, 1986]:
},|))(),(,{( XxxhxgxA AA
where,
]1,0[: XgA and ]1,0[: XhA , with the condition ,1)()(0 xhxg AA for all Xx .The numbers )(xgA and )(xhA denotes the degree of membership and the degree of non membership of x to A, respectively. For all IFS A in X we call the intutionistic index of an element Xx in A the following expression:
).()(1)( xhxgx AAA
we consider )(xA as a hesitancy degree of x to A.
Evidently, for all .
8
Evidently, 1)(0 xA for all Xx .1.2 Fuzzy Entropy
In fuzzy set theory, the entropy is measure of fuzziness which expresses the amount of average ambiguity /difficulty in taking a decision weather an element belong to a set or not. A measure of fuzziness H (A) of a fuzzy set A should have the at least the following four properties .
FE1 (Sharpness): H (A) is minimum if and only if A is crisp set, that is, A(X) =0 or 1 for all .Xx
FE2 (Maximality): H (A) is maximum if and only if A is most fuzzy set, that is, A(X) =0.5 for all .Xx
FE3 (Resolution): H (A)H(A*),where A* is the sharpened version of A.
FE4(Symmetry):H(A)=H(�),where � is the complement set of A , that is ,�(x)=1- A(x) for all .Xx
9
[Ebanks, 1993] proposed one more axiom as essential condition for validity of a measure of fuzzy entropy
FE5 (Valuation): )()()()( BHAHBAHBAH .
[DeLuca and termini, 1971] introduced a measure of fuzziness analogous to the information theoretic entropy of [Shannon, 1948] as
n
iiAiAiAiA xxxx
nAH
1
))(1log())(1()(log)(1
)( .
2. Special T-Norm
The truth table of the classical binary conjunction ^ is given in Table
1. In many-valued logic we extend the classical binary conjunction to
the unit interval as a ]1,0[]1,0[ 2 mapping as follows:
Definition 2.1 [Schweizer and Sklar, 1960] A mapping ]1,0[]1,0[: 2 C is a
From the above notions, a list of essential properties for measure of intutionistic fuzziness follows.P1 Sharpness: We have, 0)( fd i.e. 0),( AA hgRT iff. },1,0{)(),( XhXg AA i.e.
Ag and Ah are sharp.
P2 Maximality: ),(),()( AA hgRAA Thgdfd attains its maximum value only
i.e f and f can exchange intutionistic fuzziness at various points
without affecting the sum of the degrees of intutionistic fuzziness.
15
Lemma 3.1 Let }...,,,{ 21 nxxxX . A measure of fuzziness ]1,0[]1,0[]1,0[: XXd
satisfies valuation property iff. there exists a map ]1,0[]1,0[]1,0[: such that
.]1,0[]1,0[))(),(()(
1
XXn
iiAiA fxhxgfd
(2)
Lemma 3.2 Let }...,,,{ 21 nxxxX and suppose that ]1,0[]1,0[]1,0[: XXd is
given by (2) for some ]1,0[]1,0[]1,0[: . Then
(a) d satisfies P1 iff. 0)1,0()0,1( and ).1,0(,0),( yxyx
(b) d satisfies P2 iff. yx ].1,0[, yx(b) d satisfies P2 iff. yx ].1,0[, yx
(c) d satisfies P3 iff. ),(),( txzy , if ,tzyx for all x, y, z, t [0, 1].
(d) d satisfies P4 iff. ),(),( xyyx ].1,0[, yx
Theorem 3.1 Let }...,,,{ 21 nxxxX and suppose that ]1,0[]1,0[]1,0[: XXd . Then
d satisfies P1-P5, iff. d is given by (2) for some map ]1,0[]1,0[]1,0[:
satisfying conditions (a)-(d) if lemma 3.2.
16
Here, it is evident that the desirable function ]1,0[]1,0[]1,0[: in lemma
3.1 and lemma 3.2 is a restricted equivalence function. Thus, the knowledge of a restricted equivalence function gives an entropy measure of intutionistic fuzzy sets.
In earlier studies, entropy measures of intutionstic fuzzy sets have been defined using certain set of axioms. A well accepted set of axioms to define an entropy measure of intutionistic fuzzy set given by [Szmidt and Kacprzyk, 2001].and Kacprzyk, 2001].
17
[Szmidt and Kacprzyk, 2001] suggested an axiomatic definition of entropy of IFE :
Definition 3.1[Szmidt and Kacprzyk, 2001] An entropy on ),(XIF the set of
all intutionistic fuzzy sets on X is a real valued functional ],1,0[)(: XIFE
satisfying the following axiomatic requirements:IFE1: 0)( AE iff A is a crisp set, i.e. 0)( iA xg or 1)( iA xg for all .Xxi
IFE2: 0)( AE iff )()( iAiA xhxg for all .Xxi
IFE3: )()( BEAE if BA i.e. )()( iBiA xgxg and ),()( iBiA xhxh for )()( iBiB xhxg IFE3: )()( BEAE if BA i.e. )()( iBiA xgxg and ),()( iBiA xhxh for )()( iBiB xhxg
or )()( iBiA xgxg and ),()( iBiA xhxh for )()( iBiB xhxg for any .Xxi
IFE4: ).()( cAEAE
In the next section, some examples of restricted equivalence and hence new measures of intutionistic fuzzy entropy have been presented.
18
4.Some examples of REF’s and new measures of intutionistic fuzzy entropy
Some examples of REF are given as follows:
1. )1(loglog),(1 yxyx
yy
yx
xxyx
2. )1(log
)1()1(log
)1(),(
xyxyyxyxyx
19
2.2
)1(log
2
)1(
2
)1(log
2
)1(),(2
xyxyyxyxyx
3.
1
4
1
4
1
)12(
1),(3
xySin
yxSinyx
4.
1
4
1
4
1
)12(
1),(4
xyCos
yxCosyx
5.
1
42
)12(
1),(5
yxCosyx
6.
12
11
2
1
)1(
1),( 2
1
2
11
6
yxyx
eyx
eyx
eyx
7. )1(2log)1(
1),( 11
7 yxyxyxyx
20
)1(
8.
)1(
1),(
||1
8
e
eyx
yx
Measures of intutionistic fuzzy entropy corresponding to REF’s ),(1 yx
to ),(7 yx had already been studied as listed in the following table:
REF Intutionistic fuzzy entropy Author (s)
),(1 yx
n
i iAiA
iAiA
iAiA
iAiA xhxg
xhxh
xhxg
xgxg
nAE
11 )()(
)(log)(
)()(
)(log)(
1)(
)( iA x
[Vlachos and Sergiadas 2007]
),(2 yx
n xhxgxhxg )(1)()(1)(1 [Zhang
21
),(2 yx
n
i
iAiAiAiA xhxgxhxg
nAE
12 2
)(1)(log
2
)(1)(1)(
2
)(1)(log
2
)(1)( iAiAiAiA xgxhxgxh
[Zhang and Jiang, 2008]
),(3 yx
n
i
iAiA xhxgSin
nAE
13 2
)(1)(
12
11)(
2
)(1)( iAiA xgxhSin
[Ye,2010]
),(4 yx
n
i
iAiA xhxgCos
nAE
14 2
)(1)(
12
11)(
2
)(1)( iAiA xgxhCos
[Ye,2010]
),(5 yx
n
i
iAiA xhxgCos
nAE
15 1
2
)()(2
12
11)(
[Wei et al., 2012]
),(6 yx
2
)(1)(
16 2
)(1)(
)1(
11)(
iAiA xhxgn
i
iAiA exhxg
enAE
[Verma and Sharma,
22
2
)(1)(
2
)(1)(iAiA xgxh
iAiA exgxh
Sharma, 2013]
),(7 yx ))((2)()()()(log)1(
1)( 11
7 iAiAiAiAiA xxhxgxhxgn
AE
.1,0
[Verma and Sharma, 2014]
Now, we propose a new measure of intutionistic fuzzy entropy using REF ),(8 yx
n
i
xhxg iAiAeen
AE1
|)()(|1 1)1(
1)(
(4)
In order for (4), to be qualified as a suitable measure of intutionistic fuzzy entropy, it must satisfy the set IFE1-IFE4 of axiomatic requirements. In this paper, it has been proved that (4) satisfies IFE1-IFE4.
23
paper, it has been proved that (4) satisfies IFE1-IFE4.
5. Concluding Remarks
In this communication, the study of intutionistic fuzzy information measure has been done in light of a newly defined special T-Norm and restricted equivalence function. It has been established that the knowledge of a REF give rise to a new intutionistic information measure. Further, a new intutionistic fuzzy information measure have been proposed. The validity of this measures has been tested axiomatically.
24
References
[1] Atanassov, K. T., Intutionistic fuzzy sets, Fuzzy Set and Systems, 20
(1986), 87-96.
[2] Bustince, H., Barrenechea, E., Pagola M., Restricted equivalence
functions, Fuzzy sets and Systems, 157 (2006), 2333 -2346.
[3] De Luca, A. and Termini, S., A definition of non-probabilistic entropy
in the settings of fuzzy set theory, Information and Control, 20(1971),
25
in the settings of fuzzy set theory, Information and Control, 20(1971),
301-312.
[4] Ebanks, B. R., On measures of fuzziness and their representations,
J.Math Anal. and Appl. ,94(1993), 24-37.
[5] Schweizer, B., and Sklar, A., Statistical metric spaces, Pacific J. of
Mathematics, 10(1960), 313-334.
[6] Shannon, C.E., The mathematical theory of communications, Bell
Syst. Tech. Journal, 27(1948), 423–467.
[7] Szmidt, E. and Kacprzyk, J., Entropy of intutionistic fuzzy sets, Fuzzy
sets and Systems, 118 (2001), 467 -477.
[8] Verma, R. K., and Sharma, B. D., Exponential entropy of intutionistic