Neutrosophic Sets and Systems, Vol.

14, 2016 35

University of New Mexico

Interval-valued Possibility Quadripartitioned Single Valued

Neutrosophic Soft Sets and some uncertainty based
measures on them
Rajashi Chatterjee1 , Pinaki Majumdar 2 , Syamal Kumar Samanta 3
1 Department of Mathematics, Visva-Bharati, Santiniketan, 731235, India. E-mail: rajashi.chatterjee@gmail.com
2 Department of Mathematics, M. U. C. Womens College, Burdwan, 713104, India. E-mail: pmajumdar@gmail.com
3 Department of Mathematics, Visva-Bharati, Santiniketan, 731235, India. E-mail: syamal123@yahoo.co.in

Abstract: The theory of quadripartitioned single valued neutro-
sophic sets was proposed very recently as an extension to the ex-
isting theory of single valued neutrosophic sets. In this paper the
notion of possibility fuzzy soft sets has been generalized into a new
concept viz. interval-valued possibility quadripartitioned single val-
ued neutrosophic soft sets. Some basic set-theoretic operations have
been defined on them. Some distance, similarity, entropy and inclu-
sion measures for possibility quadripartitioned single valued neutro-
sophic sets have been proposed. An application in a decision making
problem has been shown.

Keywords: Neutrosophic set, entropy measure, inclusion measure, distance measure, similarity measure.

1 Introduction neutrosophic set [18], the theory of quadripartitioned single val-

The theory of soft sets (introduced by D. Molodstov, in 1999)
([10],[15]) provided a unique approach of dealing with uncer-
tainty with the implementation of an adequate parameterization
technique. In a very basic sense, given a crisp universe, a soft
set is a parameterized representation or parameter-wise classifi-
cation of the subsets of that universe of discourse with respect to
a given set of parameters. It was further shown that fuzzy sets
could be represented as a particular class of soft sets when the set
of parameters was considered to be [0, 1]. Since soft sets could
be implemented without the rigorous process of defining a suit-
able membership function, the theory of soft sets, which seemed
much easier to deal with, underwent rapid developments in fields
pertaining to analysis as well as applications (as can be seen from
the works of [1],[6],[7],[12],[14],[16],[17] etc.)
On the otherhand, hybridized structures, often designed and
obtained as a result of combining two or more existing struc-
tures, have most of the inherent properties of the combined struc-
tures and thus provide for a stronger tool in handling applica-
tion oriented problems. Likewise, the potential of the theory of
soft sets was enhanced to a greater extent with the introduction
of hybridized structures like those of the fuzzy soft sets [8], in-
tuitionistic fuzzy soft sets [9], generalized fuzzy soft sets [13],
neutrosophic soft sets [11], possibility fuzzy soft sets [2], possi-
bility intuitionistic fuzzy soft sets [3] etc. to name a few.
While in case of generalized fuzzy soft sets, corresponding to
each parameter a degree of possibility is assigned to the corre-
sponding fuzzy subset of the universe; possibility fuzzy sets, a
further modification of the generalized fuzzy soft sets, character-
ize each element of the universe with a possible degree of be-
longingness along with a degree of membership. Based on Bel-
naps four-valued logic [4] and Smarandaches n-valued refined
ued neutrosophic sets [5] was proposed as a generalization of
the existing theory of single valued neutrosophic sets [19]. In
this paper the concept of interval valued possibility quadriparti-
tioned single valued neutrosophic soft sets (IPQSVNSS, in short)
has been proposed. In the existing literature studies pertaining to
a possibility degree has been dealt with so far. Interval valued
possibility assigns a closed sub-interval of [0, 1] as the degree of
chance or possibility instead of a number in [0, 1] and thus it is
a generalization of the existing concept of a possibility degree.
The proposed structure can be viewed as a generalization of the
existing theories of possibility fuzzy soft sets and possibility in-
tuitionistic fuzzy soft sets.
The organization of the rest of the paper is as follows: a cou-
ple of preliminary results have been stated in Section 2, some
basic set-theoretic operations on IPQSVNSS have been defined
in Section 3, some uncertainty based measures viz. entropy, in-
clusion measure, distance measure and similarity measure, have
been defined in Section 4 and their p roperties, applications and
inter-relations have been studied. Section 5 concludes the paper.
2 Preliminaries
In this section some preliminary results have been outlined which
would be useful for the smooth reading of the work that follows.
2.1 An outline on soft sets and possibility intu-
itionistic fuzzy soft sets
Definition 1 [15]. Let X be an initial universe and E be a set of
parameters. Let P(X) denotes the power set of X and A E.
A pair (F, A) is called a soft set iff F is a mapping of A into

R. Chatterjee, P. Majumdar and S. K. Samanta, Interval-valued Possibility Quadripartitioned Single Valued Neutrosophic Soft Sets
and some uncertainty based measures on them
 36 Neutrosophic Sets and Systems, Vol. 14, 2016

P(X). WhenP X is discrete, A is represented as,

n
The following results are due to [3]. A = i=1 hTA (xi ), CA (xi ), UA (xi ), FA (xi )i /xi , xi X.
However, when the universe of discourse is continuous, A is
Definition 2 [3]. Let U = {x1 , x2 , ..., xn } be the univer- represented as,
sal sets of elements and let E = {e1 , e1 , ..., em } be the universal A = hTA (x), CA (x), UA (x), FA (x)i /x, xX
set of parameters. The pair (U, E) will be called a soft universe.
U U
Let F : E (I I) I U where (I I) is the collection of Definition 7 [5]. A QSVNS is said to be an absolute QSVNS,
all intuitionistic fuzzy subsets of U and I is the collection of denoted by A, iff its membership values are respectively defined
U

all fuzzy subsets of U . Let p be a mapping such that p : E I U as TA (x) = 1, CA (x) = 1, UA (x) = 0 and FA (x) = 0, xX.
U
and let Fp : E (I I) I U be a function defined as
follows: Definition 8 [5]. A QSVNS is said to be a null QSVNS,
Fp (e) = (F (e)(x), p(e)(x)), where F (e)(x) = denoted by , iff its membership values are respectively defined
(e (x) , e (x)) xU . as T (x) = 0, C (x) = 0, U (x) = 1 and F (x) = 1, xX
Then Fp is called a possibility intuitionistic fuzzy soft set (PIFSS
in short) over the soft universe (U, E). For each parameter ei , Definition 9 [5]. Let A and B be two QSVNS over X.
Fp (ei ) cann
be represented as: Then the following operations can be defined:
Containment: A B iff TA (x) TB (x), CA (x) CA (x),
  o
x1 xn
Fp (ei ) = F (ei )(x1 ) , p(ei ) (x1 ) , ..., F (ei )(xn ) , p(ei ) (xn ) UA (x) UA (x) and PFnA (x) FA (x), xX.
Complement:Ac = i=1 hFA (xi ), UA (xi ), CA (xi ), TA (xi )i /xi, xi  X
Definition 3 [3]. Let Fp and Gq be two PIFSS over (U, E). Then
i.e. TAc (xi ) = FA (xi ), CAc (xi ) = UA (xi ) , UAc (xi ) = CA (xi )
the following operations were defined over PIFSS as follows:
and FAc (xi ) = TA (xi ), xi  X
Containment: Fp is said to be a possibility intuitionistic fuzzy Pn
Union: A B = <
soft subset (PIFS subset) of Gq and one writes Fp Gq if i=1
(TA (xi ) TB (xi )) , (CA (xi ) CB (xi )) , (UA (xi ) UB (xi )) ,
(i) p(e) is a fuzzy subset of q(e), for all eE,
(FA (x) FB (x)) > /xi, xi  X
(ii)F (e) is an intuitionistic fuzzy subset of G(e), for all eE. Pn
Intersection: A B = i=1 <
Equality: Fp and Gq are said to be equal and one writes Fp = Gq
(TA (xi ) TB (xi )) , (CA (xi ) CB (xi )) , (UA (xi ) UB (xi )) ,
if Fp is a PIFS subset of Gq and Gq is a PIFS subset of Fp
(FA (xi ) FB (xi )) > /xi, xi  X
Union: Fp Gq = Hr , Hr : E (I I)U I U is de-
fined by Hr (e) = (H (e) (x) , r (e) (x)), eE such that
Proposition 1[5]. Quadripartitioned single valued neutrosophic
H (e) = Atan (F (e) , G (e)) and r (e) = s (p (e) , q (e)),
sets satisfy the following properties under the aforementioned
where Atan is Atanassov union and s is a triangular conorm.
U U set-theoretic operations:
Intersection: Fp Gq = Hr , Hr : E (I I) I is
defined by Hr (e) = (H (e) (x) , r (e) (x)), eE such that
1.(i) A B = B A
H (e) = Atan (F (e) , G (e)) and r (e) = t (p (e) , q (e)),
(ii) A B = B A
where Atan is Atanassov intersection and t is a triangular norm.
2.(i) A (B C) = (A B) C
(ii) A (B C) = (A B) C
Definition 4 [3]. A PIFSS is said to be a possibility abso-
3.(i) A (A B) = A
lute intuitionistic fuzzy soft set, denoted by A1 , if A1 : E
U (ii) A (A B) = A
(I I) I U is such that A1 (e) = (F (e) (x) , P (e) (x)), c
4.(i) (Ac ) = A
eE where F (e) = (1, 0) and P (e) = 1, eE.
(ii) Ac =
(iii) c = A
Definition 5 [3]. A PIFSS is said to be a possibility null intuition- c
U (iv) De-Morgans laws hold viz. (A B) = Ac B c ;
istic fuzzy soft set, denoted by 0 , if 0 : E (I I) I U c
(A B) = Ac B
is such that 0 = (F (e) (x) , p (e) (x)), eE where
5.(i) A A = A
F (e) = (0, 1) and p (e) = 0, eE.
(ii) A A = A
(iii) A = A
2.2 An outline on quadripartitioned single valued (iv) A =
neutrosophic sets
Definition 6 [5]. Let X be a non-empty set. A quadripartitioned
neutrosophic set (QSVNS) A, over X characterizes each element
x in X by a truth-membership function TA , a contradiction-
membership function CA , an ignorance-membership function
UA and a falsity membership function FA such that for each
x  X, TA , CA , UA , FA  [0, 1]

R. Chatterjee, P. Majumdar and S. K. Samanta, Interval-valued Possibility Quadripartitioned Single Valued Neutrosophic Soft Sets
and some uncertainty based measures on them
 Neutrosophic Sets and Systems, Vol. 14, 2016 37

3 Interval-valued possibility quadripar- Definition 12. The null IPQSVNSS over (X, E) is denoted by
0 such that for each eE and xX, e (x) = h0, 0, 1, 1i and
titioned single valued neutrosophic soft 0e (x) = [0, 0]
sets and some of their properties
3.1 Operations over IPQSVNSS
Definition 10. Let X be an initial crisp universe and E be a set of
parameters. Let I = [0, 1] , QSV N S(X) represents the collec- Definition 13. Let F and G be two IPQSVNSS over the
tion of all quadripartitioned single valued neutrosophic sets over common soft universe (X, E). Some elementary set-theoretic
X , Int([0, 1]) denotes the set of all closed subintervals of [0, 1] operations on IPQSVNSS are defined as,
and (Int([0, 1]))X denotes the collection of interval valued fuzzy (i) Union: F G = H such that for each eE and xX,
subsets over X . An interval-valued possibility quadripartitioned H e (x) = hteF (x) teG (x) , ceF (x) ceG (x) , ueF (x)
e e e
single valued neutrosophic soft set (IPQSVNSS, in short) is a uG (x) , fF (x) fG (x)iand + +
mapping of the form F : E QSV N S(X) (Int([0, 1]))X e (x) = [sup (e (x) , e (x)) , sup (e (x) , e (x))].
and is defined as F (e) = (Fe , e ) , eE, where, for each xX, (ii) Intersection: Fe G = H such that for each eE and
e e e e
Fe (x) is the quadruple which represents the truth membership, xX, H e (x) = ht F (x) tG (x) , cF (x) cG (x) , uF (x)
e e e
the contradiction-membership, the ignorance-membership and uG (x) , fF (x) fG (x)iand + +
the falsity membership of each element x of the universe of dis- e (x) = [inf (e (x) , e c(x)) , infc (e (x) , e (x))].
course X viz. Fe (x) = hteF (x) , ceF (x) , ueF (x) , fFe (x)i (iii) Complement: (F ) = F such that for each eE
c

,xX and e (x) = [e (x) , e (x)]Int([0, 1]). +
If and xX, F e (x) = hfFe (x), ueF (x), ceF (x), teF (x)i and
c +
X = {x1 , x2 , ..., xn } and E = {e1 , e2 , ..., em }, an interval- e (x) = [1 e (x) , 1 e (x)]
Containment: F G if for each eE and xX, teF (x)
valued possibility quadripartitioned single valued neutrosophic (iv) e e
as, tG (x), cF (x) cG (x) , ueF (x) ueG (x) , fFe (x) fG
e e
(x)
soft set over  the soft universe   (X, E) is represented  and (x)
(x) , +
(x) +
(x).
e e e e
F (ei ) = { Fe x(x 1
1)
, ei (x1 ) , Fe x(x 2
2)
, ei (x2 ) , ...,
 i  i
xn Example 2. Consider the IPQSNSS F and G over the
Fei (xn ) , ei (xn ) } viz.
  same soft universe (X, E) defined in example 1. Then, Fc is
x1
F (ei ) = { tei (x ),cei (x ),u , [ (x ) , +ei (x1 )] , obtained as,
h F 1 F 1 eFi (x1 ),fFei (x1 )i ei 1 x1

  Fc (e1 ) = { h0.5,0.4,0.1,0.3i , [0.4, 0.5] ,
xn
..., tei (x ),cei (x ),u , [ (x ) , + ei (xn )] }, ei E,
h F n F n eFi (xn ),fFei (xn )i ei n
   
x2 x3
h0.01,0.1,0.2,0.6i , [0.7, 0.75] , h0.6,0.4,0.3,0.7i , [0.3, 0.4] }
i = 1, 2, ..., m.  
x1
Fc (e2 ) = { h0.2,0.5,0.3,0.7i , [0.8, 0.9] ,
   
x2 x3
h0.7,0.6,0.2,0.1i , [0.4, 0.55] , h0.2,0.3,0.5,0.5i , [0.6, 0.7] }
Example 1. Let X = {x1 , x2 , x3 } and E = {e1 , e2 }.
Define an IPQSVNSS over the soft universe (X, E), H = F G is obtained as,
 
X x1
F : E QSV  N S(X) (Int([0, 1]))  as, H (e1 ) = { h0.8,0.6,0.3,0.4i , [0.8, 0.85] ,
x1    
F (e1 ) = { h0.3,0.1,0.4,0.5i , [0.5, 0.6] , x2
, [0.4, 0.5] , x3
, [0.6, 0.7] }
    h0.6,0.2,0.1,0.01i h0.7,0.5,0.3,0.4i
x2 x3  
h0.6,0.2,0.1,0.01i , [0.25, 0.3] , h0.7,0.3,0.4,0.6i , [0.6, 0.7] }
x1
  H (e2 ) = { h0.7,0.6,0.3,0.2i , [0.6, 0.75] ,
x1    
F (e2 ) = { h0.7,0.3,0.5,0.2i , [0.1, 0.2] , x 2
, [0.8, 0.9] , x3
, [0.35, 0.5] }
    h0.4,0.2,0.2,0.7i h0.9,0.7,0.1,0.2i
x2 x3 G is defined as,
h0.1,0.2,0.6,0.7i , [0.45, 0.6] , h0.5,0.5,0.3,0.2i , [0.3, 0.4] } Also, the intersection
 K = F 
x1
K (e1 ) = { h0.3,0.1,0.4,0.5i , [0.5, 0.6] ,
Another IPQSVNSS G can be definedover (X, E) as    
x2 x3
, [0.25, 0.3] , , [0.4, 0.6] }

x1
G (e1 ) = { h0.8,0.6,0.3,0.4i , [0.8, 0.85] , h0.2,0.1,0.1,0.6i
 h0.5,0.3,0.4,0.6i

x1
K (e2 ) = { h0.2,0.3,0.5,0.7i , [0.1, 0.2] ,
   
x2 x3
h0.2,0.1,0.1,0.6i , [0.4, 0.5] , h0.5,0.5,0.3,0.4i , [0.4, 0.6] }    
x2 x3
, [0.45, 0.6] , , [0.3, 0.4] }
 
x1
G (e2 ) = { h0.2,0.6,0.3,0.7i , [0.6, 0.75] , h0.1,0.2,0.6,0.7i h0.5,0.5,0.3,0.6i
   
x2 x3
h0.4,0.2,0.2,0.7i , [0.8, 0.9] , h0.9,0.7,0.1,0.6i , [0.35, 0.5] } Proposition 2. For any F , G , H IP QSV N SS(X, E),
the following results hold:
Definition 11. The absolute IPQSVNSS over (X, E) is denoted 1. (i) F G = G F
by A1 such that for each eE and xX, Ae (x) = h1, 1, 0, 0i (ii) F G = G F
and 1e (x) = [1, 1] 2. (i) F (G H ) = (F G ) H
(ii) F (G H ) = (F G ) H

R. Chatterjee, P. Majumdar and S. K. Samanta, Interval-valued Possibility Quadripartitioned Single Valued Neutrosophic Soft Sets
and some uncertainty based measures on them
 38 Neutrosophic Sets and Systems, Vol. 14, 2016

3. (i) F 0 = F |ceG (x)ueG (x)| |ceF (x)ueF (x)| and |1{+

e (x)+e (x)}|
+
(ii) F 0 = 0 |1 {e (x) + e (x)}|, xX, eE. Then,

(iii) F A1 = A1 |teG (x) fG
e
(x)|.|ceG (x) ueG (x)|.|1 {+
e (x) + e (x)}|
(iv) F A
1 = F

e e
|tF (x) fF (x)|.|cP e
F (x)
e
PuF (x)|.|1 {e (x) +
+
e (x)}|
1
4. (i) Fc = F
c 1 ||X||.||E|| eE xX |tF (x) fFe (x)|.|ceF (x)
e

(ii) Ac1 = 0 ueF (x)|.|1 {+e (x) + Pe (x)}|
1 e e e
P
 c 1 ||X||.||E|| eE xX |tG (x) fG (x)|.|cG (x)
(iii) 0 = A1 e +
uG (x)|.|1 {e (x) + e (x)}|
G )c = (F )c
5. (i) (F (G )c (F ) (G )
c c
G ) = (F )
(ii) (F (G )c
(iii) (F ) = 1
Proofs are straight-forward. 1 e e e
P P
1 ||X||.||E|| eE xX |tF (x) fF (x)|.|cF (x)

ueF (x)|.|1 { +
Pe (x)P+ e (x)}| =1
1 e
||X||.||E|| |t (x) fFe (x)|.|ceF (x) ueF (x)|.|1
4 Some uncertainty-based measures on +
eE xX F
{e (x) + e (x)}| = 0
IPQSVNSS |teF (x) fFe (x)| = 0, |ceF (x) ueF (x)| = 0,

|1 {+ e (x) + e (x)}| = 0, for each xX and each eE.
4.1 Entropy measure
tF (x) = fF (x), ceG (x) = ueG (x), +
e e
e (x) + e (x) = 1, for
each xX and each eE.
Definition 14. Let IP QSV N SS(X, E) denotes the set of
all IPQSVNSS over the soft universe (X, E). A mapping
Remark
  1. In particular, from Theorem 1, it follows that,
: IP QSV N SS(X, E) [0, 1] is said to be a measure of  

A1 = 0 and 0 = 0.
entropy if it
satisfies the following properties:
(e1) Fc = (F )
with fFe (x) fG
e Proof is straight-forward.
(e2) (F ) (G ) whenever F G (x)
e e e e e e
tG (x) tF (x), uF (x) uG (x) cG (x) cF (x) and
+
e (x) + e (x) 1. 4.1.1 An application of entropy measure in decision making
(e3) (F ) = 1 iff teF (x) = fFe (x), ceF (x) = ueF (x) and problem
+
e (x) + e (x) = 1, xX and eE.
The entropy measure not only provides an all over information
Theorem 1. The mapping e : IP QSV P
1
N SS(X,P E) e
[0, 1]
about the amount of uncertainty ingrained in a particular struc-
defined as, (F ) = 1 ||X||.||E|| eE xX F|t (x)
ture, it can also be implemented as an efficient tool in decision

fFe (x)|.|ceF (x) ueF (x)|.|1 {+
e (x) + e (x)}| is an making processes. Often while dealing with a selection process
entropy
measure for IPQSVNSS. subject to a predefined set of requisitions, the procedure involves
allocation of weights in order to signify the order of preference
Proof: of the criteria under consideration. In what follows next, the
entropy measure corresponding to an IPQSVNSS has been uti-
1


(i) Fc e lized in defining weights corresponding to each of the elements
P P
= 1 ||X||.||E|| eE |f
xX F (x)
e e e
tF (x)|.|uF (x) cF (x)|.|1 + of the parameter set over which the IPQSVNSS has been defined.
1
P P{(1 ee (x)) + (1e e (x))}|
e
= 1 ||X||.||E|| eE xX |tF (x) fF (x)|.|cF (x)
The algorithm is defined as follows:
ueF (x)|.|1 {+
e (x) + e (x)}| = (F ).

and fG e Step 1: Represent the data in hand in the form of an IPQSVNSS,

(ii) Suppose that F G (x) teG (x),
e e + say F .
uG (x) cG (x) , e (x) + e (x) 1. Automatically,
Step 2: Calculate the entropy measure (F ), as defined in
+ e e e
e (x) + e (x) 1. Thus, fF (x) fG (x), tG (x) tF (x),
e
e e e e Theorem A.
uF (x) uG (x), cG (x) cF (x), e (x) e (x) ,
Step 3: For each E, assign weights F (), given by the
+ + e e e
e (x) e (x), and fG (x) tG (x), uG (x) cG (x) ,
e
+ formula,
e (x) + e (x) 1. (F ) 1
P
F () = F () , where F () = 1 ||X||.||E|| xX |tF (x)
fFe (x) fG e
(x) teG (x) teF (x), ueF (x) ueG (x)
cG (x) cF (x) ,
e e + + fF (x)|.|c +
F (x) uF (x)|.|1 { (x) + (x)}|.
e (x) e (x) , e (x) e (x) and
+ +
e (x) + e (x) 1, e (x) + e (x) 1. Step 4: Corresponding to each option xX, calculate the net
From the above relations it follows that teG (x) fG e
(x) score, defined P as,

e e e e e e
tF (x) fF (x) but tG (x) fG (x) 0, tF (x) fF (x) 0 score(xi ) = e F ().[tF (xi ) + cF (xi ) + {1 uF (xi )} +

+
(xi )+ (xi )
|teG (x) fG e
(x)| |teF (x) fFe (x)|. Similarly, {1 fF (xi )}].{ 2 }.

R. Chatterjee, P. Majumdar and S. K. Samanta, Interval-valued Possibility Quadripartitioned Single Valued Neutrosophic Soft Sets
and some uncertainty based measures on them
 Neutrosophic Sets and Systems, Vol. 14, 2016 39

Step 5: Arrange score(xi ) in the decreasing order of values. 0.99, F (l) = 0.984
Step 6: Select maxi {score(xi )}. If maxi {score(xi )} = (4) score(x1 ) = 7.193, score(x2 ) = 9.097, score(x3 ) = 8.554
score(xm ), xm X , then xm is the selected option. (5) score(x2 ) > score(x3 ) > score(x1 )
(6) x2 is the chosen model.
Theorem 2. Corresponding to each parameter E,
(F )
F () = F () is such that 0 F () 1. 4.2 Inclusion measure
Proof: Definition 15. A mapping I : IP QSV N SS(X, E)
IP QSV N SS(X, E) [0, 1] is said to be an inclusion measure
From the definition of F () and (F ), it is clear that for IPQSVNSS over the soft universe (X, E) if it satisfies the
F () 0. following properties:

Consider |t F (x) f
F (x)|.|c
F (x) u
F (x)|.|1 { +
(x) + (I1) I A1 , 0 = 0

(x)}|.
P It follows that,
(I2) I (F , G ) = 1 F G
fF (x)|.|c +
P
|tF (x) F (x) uF (x)|.|1 { (x) + (I3) if F G
H then I (H , F ) I (G , F ) and
E xX P

(x)}|

xX |tF (x) fF (x)|.|cF (x) uF (x)|.|1 I (H , F ) I (H , G )
+
{ (x) + (x)}|, whenever P||X|| 1.
1
fF (x)|.|c
P
1 ||X||.||E|| E xX F |t (x) F (x) Theorem 3. The mapping I : IP QSV N SS(X, E) [0, 1]
1
uF (x)|.|1 { (x) + (x)}| 1 ||X||.||E|| xX |t
+
P
F (x) defined as,

fF (x)|.|c
{ + 1 e
P P
F (x) u F (x)|.|1 (x) + (x)}| I (F , G ) = 1 6||X||.||E|| eE xX [|tF (x)
(F ) F () e e e
min{tF (x), tG (x)}| + |cF (x) min{cF (x), cG (x)}| + e e
(F )
F () = F () 1, for each E. |max{ueF (x), ueG (x)} ueF (x)| + |max{fFe (x), fG e
(x)}

fF (x)| + |e (x) min{e (x), e (x)}| + |+
e
e (x)
Example 3. Suppose a person wishes to buy a phone and min{+ +
e (x), e (x)}|], is an inclusion measure for IPQSVNSS.
the judging parameters he has set are a: appearance, c: cost, b:
battery performance, s: storage and l: longevity. Further suppose Proof:
that he has to choose between 3 available models, say x1 , x2 , x3
of the desired product. After a survey has been conducted by (i) Clearly, according  to the definition of the proposed
the buyer both by word of mouth from the current users and measure, I A , = 0
1 0
the salespersons, the resultant information is represented in the
form of an IPQSVNSS, say F as follows, where it is assumed (ii) From the definition of the proposed measure, it fol-
that corresponding to an available option, a higher degree of lows that,
belongingness signifies a higher degree of agreement with the I (F , G ) = 1,
concerned parameter:
P P e
min{teF (x), teG (x)}| +
eE xX [|tF (x)
e e e

x1
 |cF (x) min{c F (x), c G (x)}| + |max{ueF (x), ueG (x)}
F (a) = { h0.4,0.3,0.1,0.5i , [0.5, 0.6] , e
|max{f e e
fFe (x)| + |
u F (x)| + F (x), f G (x)} e (x)
+ + +
   
x2 x3 min{ (x), (x)}| + | (x) min{ (x), (x)}|] =
h0.8,0.1,0.0,0.01i , [0.6, 0.7] , h0.6,0.3,0.2,0.5i , [0.45, 0.5] }
e e e e e
  0, xX, eE.
x1
F (c) = { h0.8,0.1,0.1,0.2i , [0.7, 0.75] , |teF (x) min{teF (x), teG (x)}| = 0, |ceF (x)
e e e e
min{c (x), c (x)}| = 0, |max{u (x), u (x)} ueF (x)| = 0,
   
x2 x3 F G F G
h0.5,0.01,0.1,0.6i , [0.4, 0.55] , h0.7,0.2,0.1,0.1i , [0.6, 0.65] } |max{f e e
e
|
  F (x), f G (x)} f F (x)| = 0, e (x)
x1 + + +
F (b) = { h0.65,0.3,0.1,0.2i , [0.6, 0.65] , min{ e (x), e (x)}| = 0 and |e (x)min{ e (x), e (x)}| =

x2
 
x3
 0, xX, eE.
h0.8,0.2,0.1,0.0i , [0.75, 0.8] , h0.4,0.5,0.3,0.6i , [0.7, 0.8] } Now, |teF (x) min{teF (x), teG (x)}| = 0 teF (x) teG (x).
 
x1
F (s) = { h0.5,0.4,0.3,0.6i , [0.7, 0.8] , Similarly, it can be shown that, ceF (x) ceG (x), ueF (x)
    ueG (x), fFe (x) fG e
(x),
e (x) e (x) and e (x)
+
x2 x3
h0.85,0.1,0.0,0.01i , [0.8, 0.85] , h0.8,0.2,0.1,0.02i , [0.85, 0.9] } .
+
e (x), xX, eE which proves F G
 
x1
F (l) = { h0.6,0.3,0.2,0.5i , [0.45, 0.55] ,
    (iii) Suppose, F G H . Thus we have, teF (x) teG (x)
x2 x3
h0.75,0.3,0.3,0.2i , [0.67, 0.75] , h0.75,0.3,0.2,0.2i , [0.7, 0.75] } tH (x), cF (x) cG (x) ceH (x), ueF (x) ueG (x) ueH (x),
e e e

fFe (x) fG e
(x) fH e
(x),
e (x) e (x) e (x) and

+ + +
Following steps 2-6, we have the following results: e (x) e (x) e (x) for all xX and eE.
I (H , F ) I (G , F ).
(2) (F ) = 0.982 In an exactly analogous manner, it can be shown that,
(3) F (a) = 0.984, F (c) = 0.983, F (b) = 0.988, F (s) = I (H , F ) I (H , G ). This completes the proof.

R. Chatterjee, P. Majumdar and S. K. Samanta, Interval-valued Possibility Quadripartitioned Single Valued Neutrosophic Soft Sets
and some uncertainty based measures on them
 40 Neutrosophic Sets and Systems, Vol. 14, 2016

Example 4. Consider IPQSVNSS F , G in Example 1,

4.4 Similarity measure
then I (F , G ) = 0.493.
Definition 19. A mapping s : IP QSV N SS(X, E)
4.3 Distance measure IP QSV N SS(X, E) R+ is said to be a quasi-
similarity measure between IPQSVNSS if for any
Definition 16. A mapping d : IP QSV N SS(X, E) F , G , H IP QSV N SS(X, E) it satisfies the following
IP QSV N SS(X, E) R+ is said to be a distance measure be- properties:
tween IPQSVNSS if for any F , G , H IP QSV N SS(X, E) (s1) s (F , G ) = s (G , F )
it satisfies the following properties: (s2) 0 s (F , G ) 1 and s (F , G ) = 1 F = G
(d1) d (F , G ) = d (G , F ) In addition, if it satisfies
(d2) d (F , G ) 0 and d (F , G ) = 0 F = G (s3) if F G H then s (F , H ) s (F , G ) s (G , H )
(d3) d (F , H ) d (F , G ) + d (G , H ) then it is known as a similarity measure between IPQSVNSS.
In addition to the above conditions, if the mapping d satisfies the
condition Various similarity measures for quadripartitioned single
(d4) d (F , G ) 1, F , G IP QSV N SS(X, E) valued neutrosophic sets were proposed in [5]. Undertaking a
it is called a Normalized distance measure for IPQSVNSS. similar line of approach, as in our previous work [5] we propose
a similarity measure for IPQSVNSS as follows:
Theorem 4. The mapping dh : IP QSV N SS(X, E)
IP QSV N SS(X, P E) P R+ defined as, Definition 20. Consider F , G IP QSV N SS(X, E). Define
e e e F,G
dh (F , G ) = eE xX (|tF (x) tG (x)| + |cF (x) functions i,e : X [0, 1], i = 1, 2, .., 5 such that for each
e e e e e
cG (x)| + |uF (x) uG (x)| + |fF (x) fG (x)| + |e (x) xX, eE
F,G
+ +
e (x)|+|e (x)e (x)|) is a distance measure for IPQSVNSS. 1,e (x) = |tG (x) tF (x)|
e e

It is known as the Hamming Distance. F,G e e

2,e (x) = |fF (x) fG (x)|
F,G e e
3,e (x) = |cG (x) cF (x)|
Proofs are straight-forward. F,G
4,e (x) = |ueF (x) ueG (x)|
F,G
Definition 17. The corresponding Normalized Hamming 5,e (x) = |e (x) e (x)|
F,G
distance for IPQSVNSS is defined as dN h (F , G ) = 6,e (x) = |+ e (x) e (x)|
+
1
d
6||X||.||E|| h (F , G ), where ||.|| denotes the cardinality Finally, define a mapping s : IP QSV N SS(X, E)
of a set. IP QSV N SS(X, E) R+ as, s (F , G ) = 1
1
P P P6 F,G
6||X||.||E|| eE xX i=1 i,e (x)
Theorem 5. The mapping dE : IP QSV N SS(X, E)
IP QSV N SS(X,P E) P R+ defined as, Theorem 6. The mapping s (F, G) defined above is a
e e 2 e
dE (F , G ) = eE xX {(tF (x) tG (x)) + (cF (x) similarity measure.
ceG (x))2 + (ueF (x) ueG (x))2 + (fFe (x) fG e
(x))2 + (e (x)
2 + + 2 12
e (x)) + (e (x) e (x)) } is a distance measure for Proof:
IPQSVNSS. It is known as the Euclidean Distance.
(i) It is easy to prove that s(F , G ) = s(G , F ).
Proofs are straight-forward.
(ii) We have, teF (x), ceF (x), ueF (x), fFe (x)[0, 1] and
F,G
Definition 18. The corresponding Normalized Hamming e (x), e (x)Int([0, 1]) for each xX, eE. Thus, 1,e (x)
N e e
distance for IPQSVNSS is defined as dE (F , G ) = attains its maximum value if either one of tF (x) or tG (x) is equal
1
6||X||.||E|| dE (F , G ). to 1 while the other is 0 and in that case the maximum value is 1.
Similarly, it attains a minimum value 0 if teF (x) = teG (x). So, it

Proposition 3. F G H iff follows that 0 1,e F,G
(x) 1, for each xX. Similarly it can be
(i) dh (F , H ) = dh (F , G ) + dh (G , H ) F,G
shown that i,e (x), i = 2, ..., 6 lies within [0, 1] for each xX.
(ii) dN N N
h (F , H ) = dh (F , G ) + dh (G , H ) So, P
6 F,G
0 i=1 i,e (x) 6
Proofs are straight-forward. P P Pn F,G
0 eE xX i=1 i,e (x) 6||X||.||E||
Example 5. Consider the IPQSVNSS given in Example 1. which implies 0 s(F , GPn 1.
)
The various distance measures between the sets are obtained Now s(F , G ) = 1 iff i=1 i,e (x) = 0 for each xX, eE
e e e
N
as, d (F , G ) = 5.29, d (F , G ) = 0.882,d (F , G ) = t F (x) = t G (x), cF (x) = ceG (x), ueF (x) = ueG (x),
h h E
N
4.387, dE (F , G ) = 0.731

R. Chatterjee, P. Majumdar and S. K. Samanta, Interval-valued Possibility Quadripartitioned Single Valued Neutrosophic Soft Sets
and some uncertainty based measures on them
 Neutrosophic Sets and Systems, Vol. 14, 2016 41

fFe (x) = fG
e
(x) and + +
e (x) = e (x), e (x) = e (x) , for all Theorem 7. s (F , G ) is a similarity measure.
xX, eE i.e.. iff F , G .
Proof is similar to that of Theorem 6.
(iii) Suppose F G H . then, we have, teF (x) teG (x)
tH (x), cF (x) cG (x) ceH (x), ueF (x) ueG (x) ueH (x),
e e e
Remark 3. s (F , G ) is the weighted similarity measure
fFe (x) fG e
(x) fH e
(x),
e (x) e (x) e (x) and between any two IPQSVNSS F and G .
+ + +
e (x) e (x) e (x) for all xX and eE. Con-
F,G F,G
sider 1,e (x) and 2,e (x). Since teF (x) teG (x) holds, 4.4.1 Allocation of entropy-based weights in calculating
it follows that, |tG (x) teF (x)| |teH (x) teF (x)|
e
weighted similarity
F,G F,H
1,e (x) 1,e (x). Similarly it can be shown that
F,G F,H It was shown in Section 4.1.1 how entropy measure could be
i,e (x) i,e (x), for i = 3, 5, 6 and all xX. Next, implemented to allocate specific weights to the elements of the
F,G
consider 2,e (x). parameter set. In this section, it is shown how the entropy-based
e e e
Since, fF (x) fG (x) fH (x), it follows that weights can be implemented in calculating weighted similarity.
fF (x) fG (x) fF (x) fH (x) where fFe (x) fG
e e e e e
(x) 0, Consider an IPQSVNSS F defined over the soft universe
e e e e e e
fF (x)fH (x) 0. Thus, |fF (x)fG (x)| |fF (x)fH (x)| (X, E). Let F (e)[0, 1] be the weight allocated to an element
F,G F,H
3,e (x) 3,e (x). eE, w.r.t. the IPQSVNSS F .
F,G F,H Define F () as before, viz.
Also, it can be shown that 4,e (x) 4,e (x) respectively for
(F ) 1
P
each xX. Pn F () = F () , where F () = 1 ||X||.||E|| xX |tF (x)
P P F,G
Thus, we have, eE xX i=1 i,e (x) +
fF (x)|.|cF (x) uF (x)|.|1 { (x) + (x)}|
P P Pn F,H
eE xX
i=1 i,e (x) Consider any two IPQSVNSS F , G IP QSV N SS(X). Fol-
1 6||X||.||E||1
P P Pn F,H
lowing Definition C, the weighted similarity measure between
eE xX i=1 i,e (x)
1
P P P n F,G these two sets can be defined
P as
1 6||X||.||E|| eE xX i=1 i,e (x) eE (){
P P6
F,G (x)}
s (F , G ) = 1 xX
6||X||.||E||
P i=1 i
() , where
s (F , H ) s (F , G ) eE
(G )
In an analogous manner, it can be shown that () = F ()+ 2
G ()
, and G () = G () is the weight
s (F , H ) s (G , H ). Thus, we have, s (F , H ) allocated to the parameter E w.r.t. the IPQSVNSS G .
s (F , G ) s (G , H ) From previous results clearly, F (), G ()[0, 1]
()[0, 1].
Remark 2. s(A1 , 0 ) = 0.
Example 6. Consider F , G IP QSV N SS(X) as de-
Proof : fined in Example 1. Then s (F , G ) = 0.738. Also, F (e1 ) =
0.983, G (e1 ) = 0.987, F (e2 ) = 0.993, G (e2 ) = 0.988,
For each xX and eE, which gives, (e1 ) = 0.985, (e2 ) = 0.991 which finally yields
, 1 ,0
= s (F , G ) = 0.869.
A A
1 1 0 (x) = |te (x) teA (x)| = 1, 2 (x)

0 1
|fAe (x) fe (x)| = 1
1
0
,
A
3 1 0 (x) = |ce (x) ceA (x)| = 1, 4
1 ,0
A
(x) = 5 Relation between the various uncer-

0 1
|ueA (x) ue (x)| = 1
1
0
tainty based measures
,
A ,
A
5 1 0 (x) = |
e (x) e (x)| = 1, 6 1 0 (x) = Theorem 8. s1 (F , G ) = 1 dN (F , G ) is a similarity
+ + d h
|e (x) e (x)| = 1 measure.
P6
A ,
which yields eE xX i=1 i 1 0 (x) = 6||X||.||E||
P P
P6 1 ,0
A
s(A1 , 0 ) = 1 6||X||.||E||
1
(x) = Proof:
P P
eE xX i=1 i
0.
(i) dN N 1 1
h (F , G ) = dh (G , F ) sd (F , G ) = sd (G , F )
N 1
Definition 21. Suppose F , G IP QSV N SS(X, E). (ii) 0 dh (F , G ) 1 0 sd (F , G ) 1
1 N
Consider functions i,e F,G
: X [0, 1], i = Also, sd (F , G ) = 1 dh (FN, G ) = 0 F N= G .
Whenever F G H , dh (F , H ) = dh (F , G ) +
1, 2, .., 5 as in Definition 1. Define a mapping s : (iii) N
IP QSV N SS(X, E) IP QSV N SS(X, E) R as, 1h + d (G , H ). Thus,
P
eE
P
xX
P6
i=1
F,G
(e)i,e (x) sd (F , G ) s1d (F , H ) = 1 dN h (F , G ) 1 +
s (F , G ) = 1 P
6||X||.||E|| eE (e) , where (e) is dN (F , H ) = dN (F , H ) dN (F , G ) = dN (G , H )
h h h h
the weight allocated to the parameter eE and (e)[0, 1], for 0, from property of distance measure.
each eE. s1d (F , H ) s1d (F , G ).
Similarly, it can be shown that, s1d (F , H ) s1d (G , H ).

R. Chatterjee, P. Majumdar and S. K. Samanta, Interval-valued Possibility Quadripartitioned Single Valued Neutrosophic Soft Sets
and some uncertainty based measures on them
 42 Neutrosophic Sets and Systems, Vol. 14, 2016

Hence, s1d (F , H ) s1d (F , G ) s1d (G , H ). |(F ) (G )| is a distance measure.

Remark 4. For any similarity measures (F , G ) , 1s (F , G ) Proof:

may not be a distance measure.
(i) |(F ) (G )| = |(G ) (F )|
1
Theorem 9.s2d (F , G ) = 1+dh (F ,G )
is a similarity measure. (ii) |(F ) (G )| 0 and in particular, |(F ) (G )| =
0 (F ) = (G ) (F ) (G ) and
Proof: (F ) (G ) F = G
(iii) Triangle inequality follows from the fact that,
(i) dh (F , G ) = dh (G , F ) s2d (F , G ) = s2d (G , F ) |(F ) (H )| |(F ) (G )| + |(G ) (H )|
(ii) dh (F , G ) 0 0 s2d (F , G ) 1. Also, for any F , G , H IP QSV N SS(X, E).
s2d (F , G ) = 1 dh (F , G ) = 0 F = G .
(iii) dh (F , H ) = dh (F , G ) + dh (G , H ) whenever
H
F G . 6 Conclusions and Discussions
dh (F , H ) dh (F , G ) and dh (F , H ) dh (G , H ).
1 1
1+dh (F ,H )
1+dh (F ,G )
s2d (G , F ) s2d (F , G ). In this paper, the concept of interval possibility quadripartitioned
Similarly, it can be shown that, s2d (G , F ) s2d (G , H ). single valued neutrosophic sets has been proposed. In the present
set-theoretic structure an interval valued gradation of possibil-
Corollary 1. s3d (F , G ) = 1+dN (F 1
is a similarity ity viz. the chance of occurrence of an element with respect to
,G )
measure.
h
a certain criteria is assigned and depending on that possibility of
occurrence the degree of belongingness, non-belongingness, con-
Proofs follow in the exactly same way as the previous the- tradiction and ignorance are assigned thereafter. Thus, this struc-
orem. ture comes as a generalization of the existing structures involv-
ing the theory of possibility namely, possibility fuzzy soft sets
Remark 5. For any similarity measure s (F , G ) , s(F1,G ) 1 and possibility intuitionistic fuzzy soft sets. In the present work,
may not be a distance measure. the relationship between the various uncertainty based measures
have been established. Applications have been shown where the
Theorem 10 Consider the similarity measure s (F , G ). entropy measure has been utilized to assign weights to the ele-
s (F , F G )is an inclusion measure. ments of the parameter set which were later implemented in a
decision making problem and also in calculating a weighted sim-
Proof: ilarity measure. The proposed theory is expected to have wide
applications in processes where parameter-based selection is in-
(i) Choose F = A1 and G = 0 . Then, s (F , F G ) = volved.

s(A1 , 0 ) = 0, from previous result.
(ii) s (F , F G ) = 1 F = F G F G .
(iii) Let F G H . Then, s (F , H ) s (F , G ) and 7 Acknowledgements

s (F , H ) s (G , H ) hold. Consider s (F , H )
s (F , G ). From commutative property of similarity measure, The research of the first author is supported by University JRF
it follows that, s (H , F ) s (G , F ) s (H , H F ) (Junior Research Fellowship).

s (G , G F ). Similarly, s (H , H F ) s (F , F G ). The research of the third author is partially supported by the Spe-
cial Assistance Programme (SAP) of UGC, New Delhi, India

Theorem 11.1 d (F , F G ) is an inclusion measure. [Grant no. F 510/3/DRS-III/(SAP-I)].
h

Proof follows from the results of Theorem 8 and Theorem

10. References
[1] H. Aktas and N. Cagman Soft sets and soft groups, Information Sciences,
Theorem 12. 1+dh (F1,F G
) and 1+dN
1
are in- 177 (2007), 27262735.
h (F ,F G )
clusion measures. [2] S. Alkhazaleh, A. R. Salleh and Nasruddin Hassan Possibility Fuzzy Soft
Set, Advances in Decision Sciences, 2011 (2011), 1-18.
Proofs follow from Theorem 9,Corollary 1 and Theorem [3] M. Bashir, A. R. Salleh and S. Alkhazaleh Possibility Intuitionistic Fuzzy
Soft Set, Advances in Decision Sciences, 2012 (2012), 1-24.
10.
[4] N. D. Belnap Jr. A useful four valued logic, Modern Uses of Multiple-
Valued Logic, 1904 (1977), 9-37.
Theorem 13. Let e : IP QSV N SS(X, E) [0, 1] be a
[5] R. Chatterjee, P. Majumdar and S. K. Samanta On some similarity mea-
. Then
measure of entropy such that (F ) (G ) F G sures and entropy on quadripartitioned single valued neutrosophic sets,
Journal of Intelligent and Fuzzy Systems, 30(4) (2016), 2475-2485.

R. Chatterjee, P. Majumdar and S. K. Samanta, Interval-valued Possibility Quadripartitioned Single Valued Neutrosophic Soft Sets
and some uncertainty based measures on them
 Neutrosophic Sets and Systems, Vol. 14, 2016 43

[6] D. Chen, E. C. C. Tsang, D. S. Yeung and X. Wang The parametrized
reduction of soft sets and its applications, Computers and Mathematics
with Applications, 49 (2009), 757-763.
[7] S. Das and S. K. Samanta Soft real sets, soft real numbers and their prop-
erties, Journal of Fuzzy Mathematics, 20(3) (2012), 551-576.
[8] P. K. Maji, R. Biswas and A. R. Roy Fuzzy soft sets, Journal of fuzzy
mathematics, 9(3) (2001), 589-602.
[9] P. K. Maji, R. Biswas and A. R. Roy Intuitionistic fuzzy soft sets, Journal
of fuzzy mathematics, 9(3) (2001), 677-692.
[10] P. K. Maji, R. Biswas and A. R. Roy Soft set theory, Computers and
Mathematics with Applications,45 (2003), 555-562.
[11] P. K. Maji Neutrosophic soft set, Annals of Fuzzy Mathematics and Infor-
matics, 5(1) (2013), 157-168.
[12] P. Majumdar and S. K. Samanta Similarity measureof soft sets, New
Mathematics and Natural Computation, 4(1) (2008), 1-12.
[13] P. Majumdar and S. K. Samanta Generalized fuzzy soft sets, Computers
and Mathematics with Applications, 59 (2010), 1425-1432.
[14] P. Majumdar and S. K. Samanta Softness of a soft set: soft set entropy,
Annals of Fuzzy Mathematics with Informatics, 6(1) (2013), 50-68.
[15] D. Molodstov Soft set theory-First results, Computers and Mathematics
with Applications, 37 (1999), 19-31.
[16] M. M. Mushrif, S. Sengupta and A. K. Roy Texture classification using a
novel soft-set theory based classification algorithm, Springer-Verlag Berlin
Heidelberg (2006), 246-254.
[17] D. Pei and D. Miao From soft sets to information systems, Proceedings of
Granular Computing 2 IEEE (2005), 617-621.
[18] F. Smarandache n-valued Refined Neutrosophic Logic and Its Applications
to Physics, arXiv preprint arXiv:1407.1041 (2014).
[19] H. Wang, F. Smarandache, Y. Zhang and R Sunderraman Single Valued
Neutrosophic Sets, Multispace and Multistructure, 4 (2010), 410-413.

Received: November 15, 2016. Accepted: November 22, 2016

R. Chatterjee, P. Majumdar and S. K. Samanta, Interval-valued Possibility Quadripartitioned Single Valued Neutrosophic Soft Sets
and some uncertainty based measures on them