Published Paper of International Journal of Computational Science and Information Technology (IJCSITY)

International Journal of Computational Science and Information Technology (IJCSITY) Vol. 9, No. 4, November 2021
DOI : 10.5121/ijcsity.2021.9401 1
FUZZY ROUGH INFORMATION MEASURES AND
THEIR APPLICATIONS
Seema Singh1
, D.S. Hooda2
, S.C. Malik3
1&3
Department of Statistics, M.D. University, Rohtak-124001, Haryana, India
2
Honorary Professor of Mathematics, G.J. University of Science and
Technology, Hisar- 125001, Haryana, India
Corresponding Author – D.S. Hooda
ABSTRACT
The degree of roughness characterizes the uncertainty contained in a rough set. The rough entropy was
defined to measure the roughness of a rough set. Though, it was effective and useful, but not accurate
enough. Some authors use information measure in place of entropy for better understanding which
measures the amount of uncertainty contained in fuzzy rough set .In this paper three new fuzzy rough
information measures are proposed and their validity is verified. The application of these proposed
information measures in decision making problems is studied and also compared with other existing
information measures.
KEYWORDS AND PHRASES
Fuzzy Rough Set, Similarity Information Measure, Logarithmic Information Measure, Weighted
Information Measure and Decision Making Problem
1. INTRODUCTION
The extension of crisp set theory to fuzzy set and rough set theories was developed by Zadeh [26]
and Pawlak [16] respectively. People use to compare rough set with that of fuzzy set, but both the
notions in aims and objectives are different. However, there is no sense to compare which one is
better or more useful than other. Rough set theory has its own importance in artificial intelligence
and in cognitive sciences, particularly in the areas of pattern recognition, machine learning,
inductive reasoning, knowledge acquisition, etc.
The concept of rough set theory sometimes overlaps with Dempster-Shafer theory [23] of
evidence. But the main difference between these is that the main tool in Dempster-Shafer theory
is a belief function, where in rough set theory, lower and upper approximations sets are used. No
preliminary and additional information is needed in rough set theory like membership grade in
fuzzy set theory, probability distribution in statistics and basic probability assignment in
Dempster-Shafer theory.
A rough set deals with incomplete information where a fuzzy set deals with vagueness, so it is
interesting to know how to handle the real life problems having both incomplete information as
well as vagueness in data. Thus, to handle such kind of situation Nakamura [13] and Dubois and
Prade [4] introduced the concept of fuzzy rough set, which was called as a hybrid model of fuzzy
and rough sets. Thereafter, Nanda and Majumdar [14] widely used this concept in the
development of their research. Banerjee and Pal [1] studied the roughness of fuzzy set in 1996
and Verbiest [25] worked on fuzzy rough and evolutionary approaches to instance selection.

2
Jabar and Rashid [9] combined fuzzy rough set with salient features for human resource
management classification.
Some trigonometric information measures for fuzzy rough set and their applications in medical
diagnosis were studied by Sharma and Gupta [21]. Similarity and distance information measures
on fuzzy rough set with their applications were described by Sharma et al. [19]. Recently, a new
approach to rough set based on remote system was introduced by Sun et al. [24]. Zhan et al. [28]
also studied covering based multi-granulation (I,T)- fuzzy rough set models and their applications
in multi-attribute group decision-making problems.
The word “entropy” was first used to measure an amount of uncertainty in probability
distribution of a random variable in an experiment by Shannon [18]. Later, non-probabilistic
entropy of a fuzzy set was proposed and described by Zadeh [27]. Fuzziness of a fuzzy set due to
ambiguity, impreciseness and vagueness can be measured by using fuzzy entropy which was
defined and characterized by De Luca and Termini [3].
Several other researchers, like Kapur [10], Liu [12], Pal and Pal [15], Kosko [11] and Gupta and
Sheoran [5] used Shannon’s entropy axioms of characterization to measure uncertainty in fuzzy
rough set. A similarity information measure between fuzzy rough set and fuzzy rough values was
defined by Chengyi et al. [2] and that was characterized by Qi & Chengyi [17] in 2008.
Logarithmic entropy for fuzzy rough set and its application in decision making was proposed by
Sharma [22]. The word entropy is a tedious word usually not understood easily, so fuzzy entropy
was replaced by fuzzy information measure. Thus, some authors have called fuzzy entropy as
fuzzy information measure. Hooda and Jain [6] in 2009 introduced three sub additive measures of
fuzzy information and studied their applications in medical and social sciences. A new
information measure of a fuzzy set was suggested and characterized by Hooda and Bajaj [7] and
called it as “useful” fuzzy information measure. Hooda and Raich [8] unified existing work of
various authors and described various generalizations of fuzzy information measures with their
applications.
In the present paper, some new logarithmic information measures for fuzzy rough values and
fuzzy rough set are proposed and their applications are studied. Basic concepts and definitions
used in the later development of the paper are described in section 2. In section 3, a new
logarithmic information measure for fuzzy rough values is defined and its validity is proved.
Another information measure for fuzzy rough set and its application with illustrations are studied
in section 4. In section 5, a weighted information measure for fuzzy rough set is discussed with
its application. Comparison of the information measure with other existing information measures
is studied in section 6. Conclusion is given in section 7 with references at the end of paper.
2. PRELIMINARIES
In this section some basic concepts and definitions used in development of the later part of the
paper are described and illustrated with examples.
Definition 2.1[ 26 ]
Let X be a non-empty universal set and A is a subset of X, then a function ],
1
,
0
[
:
)
( 
X
x
A

defines fuzzy set on X and is usually written as
)}
];
1
,
0
[
)
(
:
)
(
,
{( X
x
x
x
x
A i
i
A
i
A
i 


 

where 𝜇𝐴(𝑥) is called membership function from 𝑋 to [0, 1] with the following properties:

3
𝜇𝐴(𝑥) = {
0, 𝑖𝑓 𝑥 ∉ 𝐴 𝑡ℎ𝑒𝑟𝑒 𝑖𝑠 𝑛𝑜 𝑎𝑚𝑏𝑖𝑔𝑢𝑖𝑡𝑦
1, 𝑖𝑓 𝑥 ∈ 𝐴 𝑡ℎ𝑒𝑟𝑒 𝑖𝑠 𝑛𝑜 𝑎𝑚𝑏𝑖𝑔𝑢𝑖𝑡𝑦
0.5, 𝑤ℎ𝑒𝑡ℎ𝑒𝑟 𝑥 ∈ 𝐴 𝑜𝑟 𝑥 ∉ 𝐴 𝑡ℎ𝑒𝑟𝑒 𝑖𝑠
𝑚𝑎𝑥𝑖𝑚𝑢𝑚 𝑎𝑏𝑖𝑔𝑢𝑖𝑡𝑦
Definition 2.2[4]
Let 𝑈 be the universal set and 𝑅 be an equivalent relation on 𝑈 × 𝑈which is also known as
indistinguishable relation, then 𝑊 = (𝑈, 𝑅) is known as Pawlak approximation space. The set of
equivalent class generated by 𝑅 on 𝑈i.e.𝑈/𝑅 = {𝑋1, 𝑋2, … , 𝑋𝑚} is called knowledge or
equivalence class, [𝑥]𝑅is equivalent class of elements 𝑥 and the set of all fuzzy sets on 𝑈 is
denoted by 𝐹(𝑈). Now suppose ∀𝑋 ⊆ 𝑈, 𝑋 = {𝑥 ∈ 𝑈|[𝑥] ⊆ 𝑋}and 𝑋 = {𝑥 ∈ 𝑈|[𝑥] ∩ 𝑋 ≠
∅, then 𝑋 = (𝑋, 𝑋) is called rough set in 𝑊 and 𝑋 𝑎𝑛𝑑 𝑋 are called lower and upper
approximation of 𝑋 on W respectively.
Example2.1
Let U={1,2,3,4}be a universal set, B={{1,2},{3,4}} and X= {1,2,3}(X⊂U). Let B(X) be the
rough set of X, then B(X)={{1,2},{1,2, 3,4}}={{1,2,3},{1,2,4}},where{1,2}is the lower
approximation of X and{1,2,3,4} is upper approximation of X. {{1,2,3}, {1,2,4}} is family of all
sets containing {1,2} and {1,2,3,4}astheirlowerandupperapproximations.Thus,({1,2},{1,2,3,4})is
the rough set of {1,2,3}.
Definition 2.3[4]
Let 𝑈 be universe of discourse, 𝑅be a fuzzy relation on 𝑈 × 𝑈 and (𝑈, 𝑅) is fuzzy approximation
space. For any set 𝐴 ∈ 𝐹(𝑈),the lower and upper approximations of A namely 𝑅(𝐴) 𝑎𝑛𝑑 𝑅(𝐴)
with respect to approximation space (𝑈, 𝑅) are called fuzzy sets of 𝑈 whose membership
functions are defined by
𝑅(𝐴) =∨𝑦∈𝑈 [𝑅(𝑥, 𝑦) ∧ 𝐴(𝑦)], 𝑥 ∈ 𝑈 and
𝑅(𝐴) =∧𝑦∈𝑈 [(1 − 𝑅(𝑥, 𝑦)) ∨ 𝐴(𝑦)], 𝑥 ∈ 𝑈.
Hence, the pair (𝑅(𝐴), 𝑅(𝐴)) is defined as fuzzy rough set.
Example2.2.
Suppose 𝑅 = {𝑟1, 𝑟2, 𝑟3, 𝑟4, 𝑟5} be the set of objects and 𝐶 = {𝑐1, 𝑐2, 𝑐3, 𝑐4, 𝑐5, 𝑐6} be the
set of parameters, then fuzzy rough set (R, C) is given below:

4
Table 2.1: Fuzzy Rough Set
C
R
𝒄𝟏 𝒄𝟐 𝒄𝟑 𝒄𝟒 𝒄𝟓 𝒄𝟔
𝒓𝟏 (.1, .5) (.2, .6) (.4, .8) (.7, .9) (.6, .8) (.4, .6)
𝒓𝟐 (.2, .3) (.1, .3) (.4, .9) (.6, .8) (.5, .6) (.8, .8)
𝒓𝟑 (.1, .5) (.3, .5) (.5, .7) (.3, .8) (.6, .6) (.4, .9)
𝒓𝟒 (.5, .7) (.1, .5) (.3, .8) (.3, .7) (.1, .3) (.5, .8)
𝒓𝟓 (.4, .6) (.5, .7) (.3, .5) (.7, .9) (.4, .5) (.5, .6)
Definition 2.4[19]
A real-valued function that enumerates the similarity between two objects is defined as similarity
information measure. Actually, there is no particular definition of similarity information
measures, in general these information measures are some implication of inverse of distance
measures. Scores are given for similar quality, high score for more similar objects and low or
negative scores for dissimilar quality.
Let 𝑆 be the similarity measure on 𝐻, then the entropy corresponding to S is as follows:
𝑒(𝐹) = 𝑆(𝐹, 𝐹𝐶), ∀ 𝐹 ∈ 𝐻, (2.1)
where ‘𝑒’,the entropy on 𝐻 is the entropy generated by similarity information measure 𝑆 and
denoted by 𝑒(𝐹).
3. INFORMATION MEASURE FOR FUZZY ROUGH VALUES
In this section a new information measure for fuzzy rough values is proposed on t he lines of
other existing similarity information measures.
Definition 3.1[3]
A real valued function 𝑒: 𝐴 → [0, +∞), is a fuzzy information measure on 𝐴 if 𝑒 satisfies the
following four axioms ∀ 𝑥 & 𝑦 ∈ 𝐴:
a) 𝑒(𝑥) = 0, 𝑖𝑓 𝑥 = [0, 0]𝑜𝑟 𝑥 = [1, 1] 𝑖. 𝑒. 𝑥&𝑥 = 0 𝑜𝑟 1.
b) 𝑒(𝑥) = 𝑒(𝑥𝐶).
c) Fuzzy Information measure𝑒 assumes a unique value that is 𝑒(𝑥) = 1 for 𝑥 = [0.5, 0.5].
d) 𝑒(𝑥) ≥ 𝑒(𝑦), if 𝑦 is crisper than 𝑥, 𝑖. 𝑒. 𝑥 ≥ 𝑦 𝑓𝑜𝑟 𝑥 ≤ 0.5(𝑥 ≤ 0.5) 𝑎𝑛𝑑 𝑦 ≥ 𝑥 𝑓𝑜𝑟 𝑥 ≥
0.5(𝑥 ≥ 0.5).
In 2004 Chengyi et al. [2] defined a similarity information measure between two fuzzy rough
values which is as follows:
Let A be a fuzzy rough set and x, y are the fuzzy rough values in 𝐴, then the degree of similarity
between fuzzy rough values x and y is given by 𝑀𝑍 as
𝑀𝑍(x, y) = 1 −
1
2
(|𝑥 − 𝑦| − |𝑥 − 𝑦|). (3.1)
A similarity information measure between fuzzy rough sets and its elements was defined by Qi
and Chengyi [17] in 2008 as given below:

5
Let A be a fuzzy rough and x, y are the fuzzy rough values in 𝐴, then similarity degree between
elements x and y can be evaluated as
𝑀(𝑥, 𝑦) = 1 −
1
2
(𝜌𝑥𝑦 + 𝜎𝑥𝑦), (3.2)
where 𝜌𝑥𝑦 = |𝜌𝑥 − 𝜌𝑦|, 𝜎𝑥𝑦 = |𝜎𝑥 − 𝜎𝑦| and 𝜏𝑥 = 𝑥 − 𝑥 is called the degree of indeterminacy of
element 𝑥 ∈ 𝐴. 𝜌𝑥 = 𝑥 + 𝜏𝑥𝑥 = (1 + 𝜏𝑥)𝑥and 𝜎𝑥 = 1 − 𝑥 + 𝜏𝑥(1 − 𝑥) are called degree of
favour 𝑥 ∈ 𝐴.
Sharma et al. [20] defined entropy for fuzzy rough values corresponding to similarity information
measure (3.1) as follows:
𝑒(𝑥) = 1 −
1
2
(|2𝑥 − 1| + |2𝑥 − 1|), ∀𝑥 ∈ 𝐴, (3.3)
where A is a fuzzy rough set 𝑎𝑛𝑑 𝑥 = (𝑥, 𝑥).
Corresponding to the similarity information measure (3.3), logarithmic information measure for
fuzzy rough values is proposed as given below:
𝒆𝒍𝒐𝒈(𝒙) = log2(2 −
1
2
(|2𝑥 − 1| + |2𝑥 − 1|)) , ∀𝑥 ∈ 𝐴,
where A is a fuzzy rough set and 𝑥 = (𝑥, 𝑥). (3.4)
(3.4) is called a similarity fuzzy rough information measure which must satisfy the four axioms
given in definition (3.1) for its validity. Thus, these are verified by as lemmas 1to 4 and explained
below:
Lemma1
Let A be a fuzzy rough set then ∀𝑥 ∈ 𝐴 𝑎𝑛𝑑 𝑥 = (𝑥, 𝑥), the fuzzy rough information measure
defined in (3.4) is equal to zero for 𝑥 = [0, 0] 𝑜𝑟 𝑥 = [1,1] 𝑖. 𝑒. 𝑥 = 0 𝑜𝑟 1 𝑎𝑛𝑑 𝑥 = 0 𝑜𝑟 1.
Proof
Putting 𝑥 = [0, 0] 𝑖. 𝑒. 𝑥 = 0 𝑎𝑛𝑑 𝑥 = 0 in (3.4) we get
𝑒𝑙𝑜𝑔(𝑥) = log2 (2 −
1
2
(|2 × 0 − 1| + |2 × 0 − 1|))
⟹ 𝑒𝑙𝑜𝑔(𝑥) = log2 (2 −
1
2
(|0 − 1| + |0 − 1|))
⟹ 𝑒𝑙𝑜𝑔(𝑥) = log2 (2 −
1
2
× 2)
⟹ 𝑒𝑙𝑜𝑔(𝑥) = log2(2 − 1) = 𝑙𝑜𝑔21 = 0.
Again putting𝑥 = [1,1] 𝑖. 𝑒. 𝑥 = 1 𝑎𝑛𝑑 𝑥 = 1 in (3.4) we get
1
2
(|2 × 1 − 1| + |2 × 1 − 1|))
⟹ 𝑒𝑙𝑜𝑔(𝑥) = log2 (2 −
1
2
(|2 − 1| + |2 − 1|))

6
⟹ 𝑒𝑙𝑜𝑔(𝑥) = log2 (2 −
1
2
× 2)
⟹ 𝑒𝑙𝑜𝑔(𝑥) = log2(2 − 1) = 𝑙𝑜𝑔21 = 0.
Hence, for 𝑥 = [0, 0] 𝑜𝑟 𝑥 = [1,1] the fuzzy rough information measure (3.4) is zero.
Lemma2
Let A be a fuzzy rough set then ∀𝑥 ∈ 𝐴 𝑎𝑛𝑑 𝑥 = (𝑥, 𝑥), the fuzzy rough information measure
defined in (3.4) is equal to its compliment.
Proof
The measured (3.4) is equal to its compliment when𝑒𝑙𝑜𝑔(𝑥𝐶) = 𝑒𝑙𝑜𝑔(𝑥), for this put 𝑥 = 1 −
𝑥 𝑎𝑛𝑑 𝑥 = 1 − 𝑥in equation (3.4), we get
𝑒𝑙𝑜𝑔(𝑥𝐶) = 𝑙𝑜𝑔2 (2 −
1
2
(|2(1 − 𝑥) − 1| + |2(1 − 𝑥) − 1|))
⟹ 𝑒𝑙𝑜𝑔(𝑥𝐶) = 𝑙𝑜𝑔2 (2 −
1
2
(|2 − 2𝑥 − 1| + |2 − 2𝑥 − 1|)),
1
2
(|1 − 2𝑥| + |1 − 2𝑥|)),
1
2
(|2𝑥 − 1| + |2𝑥 − 1|)).
Clearly, 𝑒𝑙𝑜𝑔(𝑥𝐶) = 𝑒𝑙𝑜𝑔(𝑥), hence information measure (3.4) is equal to its compliment.
Thus, second condition is also satisfied.
Lemma3
Let A be a fuzzy rough set, then ∀𝑥 ∈ 𝐴 𝑎𝑛𝑑 𝑥 = (𝑥, 𝑥𝑧𝑧𝑦 ), the fuzzy rough information
measure defined in (3.4) is equal to one i.e. it assumes a unique maximum value for 𝑥 =
[0.5, 0.5].
Proof
Let us put 𝑥 = 0.5 𝑎𝑛𝑑 𝑥 = 0.5 in equation (3.4), then
1
2
(|2 × 0.5 − 1| + |2 × 0.5 − 1|)) ,
⟹ 𝑒𝑙𝑜𝑔(𝑥) = log2 (2 −
1
2
(|1.0 − 1| + |1.0 − 1|)),
⟹ 𝑒𝑙𝑜𝑔(𝑥) = log2 (2 −
1
2
(|0| + |0|)) ,
⟹ 𝑒𝑙𝑜𝑔(𝑥) = log2(2 − 0) = 𝑙𝑜𝑔2(2) = 1.
Thus, for 𝑥 = [0.5, 0.5]fuzzy rough information measure (3.4) assumes a unique maximum
value 1. Hence third property is satisfied.

7
Lemma 4
Let A be a fuzzy rough set, then ∀𝑥 ∈ 𝐴 𝑎𝑛𝑑 𝑥 = (𝑥, 𝑥), we have 𝑒𝑙𝑜𝑔(𝑥) ≥ 𝑒𝑙𝑜𝑔(𝑦) if 𝑦 is
sharper or crisper than 𝑥 i.e. 𝑦 ≤ 𝑥 𝑓𝑜𝑟 𝑥 ≤ 0.5(𝑥 ≤ 0.5) 𝑎𝑛𝑑 𝑦 ≥ 𝑥 𝑓𝑜𝑟 𝑥 ≥ 0.5(𝑥 ≥ 0.5).
Proof
For the first case let 𝑦 ≤ 𝑥 𝑓𝑜𝑟 𝑥 ≤ 0.5(𝑥 ≤ 0.5),
⟹ 𝑦 ≤ 𝑦 ≤ 𝑥 ≤ 0.5 &𝑦 ≤ 𝑥 ≤ 𝑥 ≤ 0.5,
⟹ |2𝑥 − 1| ≤ |2𝑦 − 1| 𝑎𝑛𝑑 |2𝑥 − 1| ≤ |2𝑦 − 1|,
⟹
1
2
(|2𝑥 − 1| + |2𝑥 − 1|) ≤
1
2
(|2𝑦 − 1| + |2𝑦 − 1|).
Subtracting the equation on both side from ‘2’ and taking binary logarithm we get
𝑙𝑜𝑔2 (2 −
1
2
(|2𝑥 − 1| + |2𝑥 − 1|)) ≥ 𝑙𝑜𝑔2 (2 −
1
2
(|2𝑦 − 1| + |2𝑦 − 1|)),
⟹ 𝑒𝑙𝑜𝑔(𝑥) ≥ 𝑒𝑙𝑜𝑔(𝑦).
Similarly, second case can be proved. Thus, all the four axioms given in definition (3.1) are
satisfied by fuzzy rough information measure (3.4). Hence, it is a valid information measure.
4. INFORMATION MEASURE FOR FUZZY ROUGH SETS
In the previous section we have defined an information measure for fuzzy rough values. Next,
corresponding to equation (3.4) another information measure for fuzzy rough set is proposed. Let
us consider a fuzzy rough set ‘A’, then information measure for fuzzy rough set ‘A’ is proposed
as given below:
𝐸𝑙𝑜𝑔(𝐴) =
1
𝑛
∑ 𝑙𝑜𝑔2
𝑛
𝑖=1
(2 −
1
2
(|2𝑥𝑖 − 1| + |2𝑥𝑖 − 1|)) , ∀ 𝑥𝑖 ∈ 𝐴 𝑎𝑛𝑑 𝑥𝑖 = (𝑥𝑖, 𝑥𝑖)
(4.1)
(4.1) is called as a fuzzy rough information measure and it lies in the interval [0, 1]. Larger value
of the fuzzy rough information measure indicates more uncertainty in ‘𝐴′. Obviously, four
lemmas for the validity of the information measure hold.
Next, some propositions are enumerated as
Proposition 4.1
Let 𝐴 be a fuzzy rough set,then ∀ 𝑥𝑖 ∈ 𝐴value of information measure (4.1) is equal to zero for
𝑥𝑖 = 0 𝑜𝑟 1 𝑎𝑛𝑑 𝑥𝑖 = 0 𝑜𝑟 1 𝑖. 𝑒. for A to be a crisp set.
Proposition 4.2
Let 𝐴 be a fuzzy rough set, then ∀ 𝑥𝑖 ∈ 𝐴 value of information measure (4.1) is equal to the value
of its complement, where 𝐴𝐶
= (1 − 𝑥𝑖, 1 − 𝑥𝑖).
Proposition 4.3

8
Let𝐴 be a fuzzy rough set, then ∀ 𝑥𝑖 ∈ 𝐴 value of information measure (4.1) is equal to one i.e. it
assumes a unique maximum value for 𝑥𝑖 = 0.5 𝑎𝑛𝑑 𝑥𝑖 = 0.5.
Proposition 4.4
Let 𝐴, 𝐵 be fuzzy rough sets,𝑡ℎ𝑒𝑛 ∀ 𝑥𝑖 ∈ 𝐴 𝑎𝑛𝑑 𝑥𝑗 ∈ 𝐵,value of information measure for 𝐴 is
greater than or equal to the value of information measure for 𝐵 when 𝐵 is sharper or crisper than
A that is 𝐵 ≤ 𝐴 𝑓𝑜𝑟 𝑥𝑖 ≤ 0.5(𝑥𝑖 ≤ 0.5) 𝑎𝑛𝑑 𝐵 ≥ 𝐴 𝑓𝑜𝑟 𝑥𝑖 ≥ 0.5 (𝑥𝑖 ≥ 0.5).
These propositions can be proved by following the same procedure as described in previous
section 3.
4.1.Application and Illustration
In this section application of fuzzy rough information measures (3.4) and (4.1) is illustrated with
two examples.
Example 4.1.
Let us consider the case of choice of restaurants in a particular city by the customers due to
several factors such as quality of food, service provide, behaviour of staff etc. It is clear from
customer’s behaviour that they are not happy with any one restaurant. They prefer different
restaurants on the basis of their taste. This whole situation is represented in the form of fuzzy
rough set as follows:
Suppose 𝑅 = {𝑟1, 𝑟2, 𝑟3, 𝑟4, 𝑟5} be the set of restaurants in a particular city and
𝐶 = {𝑐1, 𝑐2, 𝑐3, 𝑐4, 𝑐5, 𝑐6} be the set of attributes on basis of the taste of customers as 𝑐1 =
𝑖𝑛𝑡𝑒𝑟𝑖𝑜𝑟 𝑑𝑒𝑐𝑜𝑟𝑎𝑡𝑖𝑜𝑛, 𝑐2 = 𝑏𝑒𝑠𝑡 𝑝𝑜𝑠𝑠𝑖𝑏𝑙𝑒 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑚𝑜𝑛𝑒𝑦, 𝑐3 = 𝑙𝑜𝑐𝑎𝑡𝑖𝑜𝑛, 𝑐4 =
𝑏𝑒ℎ𝑎𝑣𝑖𝑜𝑢𝑟 𝑜𝑓 𝑠𝑡𝑎𝑓𝑓, 𝑐5 = 𝑞𝑢𝑎𝑙𝑖𝑡𝑦 𝑜𝑓 𝑓𝑜𝑜𝑑 𝑎𝑛𝑑 𝑐6 = 𝑜𝑓𝑓𝑒𝑟 𝑎𝑣𝑎𝑖𝑙𝑎𝑏𝑙𝑒.
The aim is to find out which restaurant is best in term of which attribute by using the information
given by customers. The information available is model in the form of fuzzy rough set and
represented in tabular form as given below:
C
R
𝑐1 𝑐2 𝑐3 𝑐4 𝑐5 𝑐6
𝑟1 (.1, .5) (.2, .6) (.4, .8) (.7, .9) (.6, .8) (.4, .6)
𝑟2 (.2, .3) (.1, .3) (.4, .9) (.6, .8) (.5, .6) (.8, .8)
𝑟3 (.1, .5) (.3, .5) (.5, .7) (.3, .8) (.6, .6) (.4, .9)
𝑟4 (.5, .7) (.1, .5) (.3, .8) (.3, .7) (.1, .3) (.5, .8)
𝑟5 (.4, .6) (.5, .7) (.3, .5) (.7, .9) (.4, .5) (.5, .6)
From above table we find the uncertainty among restaurants and given attribute by using
information measure for fuzzy rough set (4.1) as 𝐸𝑙𝑜𝑔(𝑅: 𝐶) = 0.626786.
Now we find out the uncertainties among restaurants and attributes by using information measure
(3.4) for fuzzy rough values which are given below in table 4.2:

9
Table 4.2: Uncertainties among Restaurants and Attributes
C
R
𝑐1 𝑐2 𝑐3 𝑐4 𝑐5 𝑐6
𝑒𝑙𝑜𝑔(𝑟1) 0.678072 0.678072 0.678072 𝟎. 𝟒𝟖𝟓𝟒𝟐𝟕 0.678072 0.847997
𝑒𝑙𝑜𝑔(𝑟2) 0.584963 0.584963 0.584963 0.678072 0.926 𝟎. 𝟒𝟖𝟓𝟒𝟐𝟕
𝑒𝑙𝑜𝑔(𝑟3) 0.678072 0.847997 0.847997 0.584963 0.847997 𝟎. 𝟓𝟖𝟒𝟗𝟔𝟑
𝑒𝑙𝑜𝑔(𝑟4) 0.847997 0.678072 0.584963 0.678072 𝟎. 𝟒𝟖𝟓𝟒𝟐𝟕 0.765535
𝑒𝑙𝑜𝑔(𝑟5) 0.847997 0.847997 0.847997 𝟎. 𝟒𝟖𝟓𝟒𝟐𝟕 0.926 0.926
Comparing the uncertainty 𝐸𝑙𝑜𝑔(𝑅: 𝐶)with those in table 4.2 and choosing the uncertainty less
than or equal to 𝐸𝑙𝑜𝑔(𝑅: 𝐶)we can conclude that restaurants 1and 5 are good in terms of attribute
𝑐4 i.e. behaviour of staff of restaurants 2 and 3 are good in terms of attribute 𝑐6 i.e. offer available
and restaurant 4 is good in terms of attribute 𝑐5 i.e. quality of food.
Example 4.2.
Covid-19 pandemic has transformed individual lives as well as social life on global scale. The
most effected sector by this virus is healthcare. Healthcare personnel have faced a significant
higher risk of infection, particularly in the early stages of the outbreak. On an average it takes 5-6
days for symptoms to appear when someone is infected with the corona virus. Symptoms of
covid-19 are divided into three categories as
Most Common Symptoms
 Dry Cough
 Tiredness
 Fever
Less Common Symptoms
 Headache
 Diarrhoea
 Sore throat
 Rash on Skin
 Discolouration of Finger
 Aches and Pain
 Conjunctivitis
 Loss of taste or Smell
Serious Symptoms
 Chest Pain
 Loss of Movement or Speech
 Shortness of breath or Difficulty in Breathing
If you have serious type of symptoms then immediate medical attention is required. For most
common and mild symptoms, people should stay at home and consult doctor for precaution and
cure.
Here we make use of fuzzy rough information measure to detect the type of symptoms in patient
for covid-19 and suggest whether patient requires immediate medical attention or not. To see the

10
application of the fuzzy rough information measure a hypothetical case study is framed as
follows:
Let }
,
,
,
{ 4
3
2
1 A
A
A
A
A  is the set of patients under study and
}
,
,
,
,
{ 5
4
3
2
1 ech
spe
of
Loss
T
breathing
in
Difficulty
T
pain
Chest
T
Headache
T
Fever
T
T 





are the symptoms of the patients. Symptoms 3
T , 4
T , 5
T are the serious type of symptoms and if
they occur then immediate treatment is required. The data for the relation between patients and
symptoms is given hypothetically as
Table 4.3: Fuzzy Rough Set for Patients and Symptoms
T
A
1
T 2
T 3
T 4
T 5
T
1
A [0.0,0.4] [0.6,0.9] [0.1,0.5] [0.3,0.7] [0.5,0.7]
2
A [0.3,0.5] [0.2,0.6] [0.2,0.3] [0.5,0.9] [0.6,0.9]
3
A [0.7,0.8] [0.0,0.9] [0.2,0.3] [0.1,0.2] [0.7,1]
4
A [0.9,1] [0.3,0.8] [0.3,0.4] [0.2,0.6] [0.5,0.5]
Using fuzzy rough information measure (3.4), the uncertainties for patients and symptoms are
given below in table 4.4:
Table 4.4: Uncertainties among Patients and Symptoms
T
A
1
T 2
T 3
T 4
T 5
T
1
A 0.4855 0.585 0.6781 0.6781 0.848
2
A 0.848 0.6781 0.585 0.6781 0.585
3
A 0.585 0.1375 0.585 0.3785 0.3785
4
A 0.1375 0.585 0.7656 0.6781 1
Fig 1: Uncertainties among Patients and Symptoms
T…
T…
T…
T…
T…
0
0.5
1
A1
A2
A3
A4
T1 T2 T3 T4 T5

11
The uncertainty for complete set using fuzzy rough information measure (4.1) is
59398
.
0
)
;
(
log 
T
A
E 0. Now, compare the uncertainty )
;
(
log T
A
E with those in table 4.4, if
uncertainties for symptoms 3
T , 4
T , 5
T together are greater than )
;
(
log T
A
E , then the patient is
suffering from serious symptoms of covid-19. From table 4.4 it is clear that patients 4
1 A
and
A
are suffering from serious symptoms and need immediate medical treatment, while patients
3
2 A
and
A are less effected by virus and can take home remedies to cure themselves.
5. WEIGHTED INFORMATION MEASURE FOR FUZZY ROUGH SET
In this section weighted information measure for fuzzy rough set is proposed corresponding to
measure (3.4) and an application is presented.
Let us consider a fuzzy rough set ′𝐴′,∀𝑥𝑖 ∈ 𝐴 and 𝑤𝑖 ∈ [0,1] is the weight for
element 𝑥𝑖 𝑜𝑓 𝐴, then weighted fuzzy rough information measure for A is given below:
𝐻𝑙𝑜𝑔(𝐴) =
∑ 𝑤𝑖𝑙𝑜𝑔2
𝑛
𝑖=1 (2 −
1
2
(|2𝑥𝑖 − 1| + |2𝑥𝑖 − 1|))
∑ 𝑤𝑖
𝑛
𝑖=1
,
∀ 𝑥𝑖 ∈ 𝐴, 𝑤𝑖 ∈ [0, 1]𝑎𝑛𝑑 𝑥𝑖 = (𝑥𝑖, 𝑥𝑖). (5.1)
(5.1) is called weighted fuzzy rough information measure. Next, we state some propositions for
the validity of (5.1).
Proposition 5.1
Let 𝐴 be a fuzzy rough set, then ∀ 𝑥𝑖 ∈ 𝐴 value of weighted information measure (5.1) is equal to
zero for 𝑥𝑖 = 0 𝑜𝑟 1 𝑎𝑛𝑑 𝑥𝑖 = 0 𝑜𝑟 1 𝑖. 𝑒. for A to be a crisp set.
Proposition 5.2
Let 𝐴 be a fuzzy rough set, then ∀ 𝑥𝑖 ∈ 𝐴 the value of weighted information measure (5.1) is
equal to the value of its complement, where 𝐴𝐶
= (1 − 𝑥𝑖, 1 − 𝑥𝑖).
Proposition 5.3
Let 𝐴 be a fuzzy rough set, then ∀ 𝑥𝑖 ∈ 𝐴 the value of weighted information measure (5.1) is
equal to one i.e. it assumes a unique maximum value for 𝑥𝑖 = 0.5 𝑎𝑛𝑑 𝑥𝑖 = 0.5.
Proposition 5.4
Let 𝐴, 𝐵 be fuzzy rough sets, then ∀ 𝑥𝑖 ∈ 𝐴 𝑎𝑛𝑑 𝑥𝑗 ∈ 𝐵, the value of weighted information
measure for 𝐴 is greater than or equal to the value of weighted information measure for 𝐵 when
𝐵 is sharper or crisper than A that is 𝐵 ≤ 𝐴 𝑓𝑜𝑟 𝑥𝑖 ≤ 0.5(𝑥𝑖 ≤ 0.5) 𝑎𝑛𝑑 𝐵 ≥ 𝐴 𝑓𝑜𝑟 𝑥𝑖 ≥
0.5 (𝑥𝑖 ≥ 0.5).
It may noted that the above proposition hold good obviously, so the weighted fuzzy information
measure is a valid measure.

12
5.1. Application for Illustration
For illustration purpose the data set of example 4.1 is considered in the following table:
C
R
𝑐1 𝑐2 𝑐3 𝑐4 𝑐5 𝑐6
𝑟1 (.1, .5) (.2, .6) (.4, .8) (.7, .9) (.6, .8) (.4, .6)
𝑟2 (.2, .3) (.1, .3) (.4, .9) (.6, .8) (.5, .6) (.8, .8)
𝑟3 (.1, .5) (.3, .5) (.5, .7) (.3, .8) (.6, .6) (.4, .9)
𝑟4 (.5, .7) (.1, .5) (.3, .8) (.3, .7) (.1, .3) (.5, .8)
𝑟5 (.4, .6) (.5, .7) (.3, .5) (.7, .9) (.4, .5) (.5, .6)
The uncertainties calculated on applying fuzzy rough information measure (3.4) are given in table
5.2.
Table 5.2: Uncertainties among Restaurants and Attributes
C
R
𝑐1 𝑐2 𝑐3 𝑐4 𝑐5 𝑐6
𝑒𝑙𝑜𝑔(𝑟1) 0.678072 0.678072 0.678072 0.485427 0.678072 0.847997
𝑒𝑙𝑜𝑔(𝑟2) 0.584963 0.584963 0.584963 0.678072 0.926 0.485427
𝑒𝑙𝑜𝑔(𝑟3) 0.678072 0.847997 0.847997 0.584963 0.847997 0.584963
𝑒𝑙𝑜𝑔(𝑟4) 0.847997 0.678072 0.584963 0.678072 0.485427 0.765535
𝑒𝑙𝑜𝑔(𝑟5) 0.847997 0.847997 0.847997 0.485427 0.926 0.926
Now, the weights corresponding to each element of table 5.1 are given below:
Table 5.3: Weights Corresponding to Each Element of Fuzzy Rough Set
0.53 0.62 0.55 0.82 0.92 0.87
0.84 0.80 0.66 0.45 0.33 0.23
0.16 0.25 0.39 0.64 0.87 0.33
0.29 0.36 0.46 0.47 0.78 0.63
0.13 0.85 0.74 0.53 0.24 0.82
Using weighted fuzzy rough information measure (5.1), the weighted uncertainties for fuzzy
rough values are
Table 5.4: Uncertainties Using Weighted Fuzzy Rough Information Measure (5.1)
C
R
𝒄𝟏 𝒄𝟐 𝒄𝟑 𝒄𝟒 𝒄𝟓 𝒄𝟔
𝒉𝒍𝒐𝒈(𝒓𝟏) 0.3594 0.4204 0.3730 0.3980 0.6239 0.7378
𝒉𝒍𝒐𝒈(𝒓𝟐) 0.4914 0.468 0.3861 0.3051 0.3056 0.1116
𝒉𝒍𝒐𝒈(𝒓𝟑) 0.1085 0.212 0.3307 0.3744 0.7378 0.1931
𝒉𝒍𝒐𝒈(𝒓𝟒) 0.2459 0.2441 0.2691 0.3187 0.3786 0.4823
𝒉𝒍𝒐𝒈(𝒓𝟓) 0.1102 0.7208 0.6275 0.2573 0.2222 0.7593

13
Weighted uncertainty for complete fuzzy rough set is 6988
.
0
)
;
(
log 
T
A
H .After assigning the
weights it can be noted that the decision regarding restaurants is changed. All the restaurants are
up to the expectations of customers except restaurants 𝑟1, 𝑟3 𝑎𝑛𝑑 𝑟5, particularly for attributes
𝑐6of restaurant 𝑟1, 𝑐5 of restaurant 𝑟3and 𝑐2, 𝑐6 of restaurant 𝑟5. Since, the uncertainties
corresponding to these attributes are greater than )
;
(
log T
A
H .
6. COMPARISON WITH OTHER FUZZY ROUGH INFORMATION MEASURES
In this section information measure (3.4) and (4.1) are compared with the existing fuzzy rough
information measures defined and studied by Sharma and Gupta [21].
Let ‘A’ be a rough set then, sine trigonometric information measures for fuzzy rough values of
set ‘A’ and for fuzzy rough set ‘A’ are written respectively as given below:
|))]
1
2
|
|
1
2
(|
2
1
1
(
2
sin[
)
(
sin 



 x
x
x
e

, for every element ,
A
x (6.1)
and |))]
1
2
|
|
1
2
(|
2
1
1
(
2
[
sin
1
)
(
1
sin 



 

i
i
n
i
x
x
n
x
E

, for the whole set A. (6.2)
Similarly, the cosine and tangent trigonometric information measures for fuzzy rough values and
for fuzzy rough set respectively are
|)]
1
2
|
|
1
2
(|
4
cos[
)
(
cos 


 x
x
x
e

A
x (6.3)
|)]
1
2
|
|
1
2
(|
4
[
cos
1
)
(
1
cos 


 

i
i
n
i
x
x
n
x
E

, for the whole set A. (6.4)
|))]
1
2
|
|
1
2
(|
2
1
1
(
4
tan[
)
(
tan 



 x
x
x
e

A
x (6.5)
and







n
i
i
i x
x
n
x
E
1
tan |))]
1
2
|
|
1
2
(|
2
1
1
(
4
tan[
1
)
(

,for the whole set A. (6.6)
The uncertainties for fuzzy rough values in example 4.1 are computed using the above
trigonometric fuzzy rough information measures as given in table 6.1.

14
Table 6.1: Uncertainties using Trigonometric FRI Measures
C
R
𝒄𝟏 𝒄𝟐 𝒄𝟑 𝒄𝟒 𝒄𝟓 𝒄𝟔
𝒆𝒔𝒊𝒏(𝒓𝟏) .8090 .8090 .8090 .5878 .8090 .9511
𝒆𝒄𝒐𝒔(𝒓𝟏) .8090 .8090 .8090 .5878 .8090 .9511
𝒆𝒕𝒂𝒏(𝒓𝟏) .5095 .5095 .5095 .3249 .5095 .7265
𝒆𝒔𝒊𝒏(𝒓𝟐) .7071 .5878 .7071 .8090 .9877 .5878
𝒆𝒄𝒐𝒔(𝒓𝟐) .7071 .5878 .7071 .8090 .9877 .5878
𝒆𝒕𝒂𝒏(𝒓𝟐) .4142 .3249 .4142 .5095 .8541 .3249
𝒆𝒔𝒊𝒏(𝒓𝟑) .8090 .9511 .9511 .7071 .9511 .7071
𝒆𝒄𝒐𝒔(𝒓𝟑) .8090 .9511 .9511 .7071 .9511 .7071
𝒆𝒕𝒂𝒏(𝒓𝟑) .5095 .7265 .7265 .4242 .7265 .4142
𝒆𝒔𝒊𝒏(𝒓𝟒) .9511 .8090 .7071 .8090 .5878 .8910
𝒆𝒄𝒐𝒔(𝒓𝟒) .9511 .8090 .7071 .8090 .5878 .8910
𝒆𝒕𝒂𝒏(𝒓𝟒) .7265 .5095 .4142 .5095 .3249 .6128
𝒆𝒔𝒊𝒏(𝒓𝟓) .9511 .9511 .9511 .5878 .9877 .9877
𝒆𝒄𝒐𝒔(𝒓𝟓) .9511 .9511 .9511 .5878 .9877 .9877
𝒆𝒕𝒂𝒏(𝒓𝟓) .7265 .7265 .7265 .3249 .8541 .8541
The uncertainties calculated using the trigonometric fuzzy rough information measures for fuzzy
rough set are ,
8136
.
0
)
(
sin 
R
E 8136
.
0
)
(
cos 
R
E and .
55862
.
0
)
(
tan 
R
E
Fig 2. Uncertainties Using Trigonometric FRI measures
Next, the uncertainties calculated using information measure (3.4) for fuzzy rough values of
example 4.1 are given below:
0
0.2
0.4
0.6
0.8
1
1.2
Uncertainties Using Trigonometric Information Measures
c1 c2 c3 c4 c5 c6

15
Table 6.2: Uncertainty Using FRI Measure (3.4)
C
R
𝑐1 𝑐2 𝑐3 𝑐4 𝑐5 𝑐6
𝑒𝑙𝑜𝑔(𝑟1) 0.678072 0.678072 0.678072 𝟎. 𝟒𝟖𝟓𝟒𝟐𝟕 0.678072 0.847997
𝑒𝑙𝑜𝑔(𝑟2) 0.584963 0.584963 0.584963 0.678072 0.926 𝟎. 𝟒𝟖𝟓𝟒𝟐𝟕
𝑒𝑙𝑜𝑔(𝑟3) 0.678072 0.847997 0.847997 0.584963 0.847997 𝟎. 𝟓𝟖𝟒𝟗𝟔𝟑
𝑒𝑙𝑜𝑔(𝑟4) 0.847997 0.678072 0.584963 0.678072 𝟎. 𝟒𝟖𝟓𝟒𝟐𝟕 0.765535
𝑒𝑙𝑜𝑔(𝑟5) 0.847997 0.847997 0.847997 𝟎. 𝟒𝟖𝟓𝟒𝟐𝟕 0.926 0.926
Now, uncertainty computed by information measure (4.1) for fuzzy rough set is𝐸𝑙𝑜𝑔(𝑅: 𝐶) =
0.626786.
Fig 3: Uncertainties Using fuzzy rough Information Measure (3.4)
From the table 6.1 and 6.2 and from the graphs it may be noted that the uncertainties calculated
on applying information measure (3.4) are much less than that of sine and cosine trigonometric
information measures studied by Sharma and Gupta [21]. Thus, our information measure is more
useful and simple than the trigonometric information measures. However, the uncertainty
calculated using sine and cosine measures are almost same.
7. CONCLUSION
As we know a right decision can change one’s life, so it is important to study decision making
methods for solving the problems of our daily life and fuzzy rough set theory is one of the
choices. Fuzzy rough set is the hybridization of rough set and fuzzy set which had been widely
used to deals with complex data containing different types of uncertainties. In this paper a
logarithmic information measure for fuzzy rough values is proposed and verified axiomatically.
Logarithmic information measure for fuzzy rough set and weighted logarithmic fuzzy rough
information measure are also defined with their application in decision making problem. The
Proposed fuzzy rough information measure is compared with other existing trigonometric fuzzy
rough information measures and it is proved that our information measure is better and simple.
0
0.2
0.4
0.6
0.8
1
elog(r1) elog(r2) elog(r3) elog(r4) elog(r5)
Uncertainties using fuzzy Information Measure
(3.4)
c1 c2 c3 c4 c5 c6

16
The research on application of fuzzy rough set theory can be extended in studying data dimension
reduction technique and decision making in design of experiment. A new form of rough set that is
hyper rough set can be defined. Hyper rough set is completely a new idea in the area of rough set
theory. In future researchers can put their attention towards defining new information measures
by combining logarithmic and trigonometric functions as combination of two functions overcome
the shortcoming of one another and provide better result than single one.
CONFLICT OF INTEREST
There is no conflict of Interest among the authors.
REFERENCES
[1] Banerjee M. and Pal S.K. Roughness of a Fuzzy Set. Informatics and Computer Sciences, 93(1996),
235-246.
[2] Chengyi Z, Pingan D. and Haiyan F. On Measures of Similarity between Fuzzy Rough Sets.
International Journal of Pure and Applied Mathematics, 10(2004), 451-460.
[3] De Luca and Termini S. A Definition of a Non-Probabilistic Entropy in the Setting of Fuzzy Sets
Theory. Fuzzy Sets and Systems, 52(1972), 305-318.
[4] Dubois D. and Prade H. Rough Fuzzy Sets and Fuzzy Rough Sets. International Journal of General
Systems, 17(1990), 191-209.
[5] Gupta P. and Sheoran A. Some New Parametric Fuzzy Entropies. International Journal of Fuzzy
Mathematics and Systems, 4(2014), 293-298.
[6] Hooda D.S. and Divya Jain. Sub additive Measures of Fuzzy Information. Journal of Reliability and
Statistical Studies, 2(2009), 39-52.
[7] Hooda D.S. and Bajaj R.K. “Useful”-Fuzzy Measure of Information, Integrated Ambiguity and
Directed Divergence. International Journal of General Systems, Taylor and Francis, First Article,
(2010), 10-12.
[8] Hooda D.S. and Raich V. Fuzzy Information Measures with Applications. Alpha Science,
International LTD., 7200, The Quorum, Oxford Business Park North Garsington Road, Oxford OX4
2JZ, U.K. ISBN- 978-1-84265-934-2(2015).
[9] Jabar A.L. and Rashid T.A. Combining Fuzzy Rough Set with Salient Features for HRM
Classification. Conference- 15th
IEEE International Conference on Computer and Information
Technology (CIT-2015), At. Liverpool, England, U.K., 244-251. DOI:
10.1109/CIT/1UCC/DASC/PICOM.2015.35.
[10] Kapur T.N. Measure of Fuzzy Information. New Delhi Mathematical Sciences of Trust Society
(1997).
[11] Kosko B. Fuzzy Entropy and Conditioning. Information Sciences, 40(1986), 165-174.
[12] Liu X.C. Entropy Distance Measure and Similarity Measure of Fuzzy Sets and Their Relations. Fuzzy
Sets and Systems, 52(1992), 305-318.
[13] Nakamura A. Fuzzy Rough Sets. Note on Multiple Valued Logic in Japan, 9(1988), 1-8.
[14] Nanda S. and Majumdar S. Fuzzy Rough Sets. Fuzzy Sets and Systems, 145(1992), 157-160.
[15] Pal N.R. and Pal S.K. Entropy-A New Definition and its Applications. IEEE Transactions on
Systems, Man and Cybernetics, 21(1999), 1260-1270.
[16] Pawlak Z. Rough Sets. International Journal of Information and Computational Sciences,11(1982),
341-356.
[17] Qi N. and Chengyi Z. A New Similarity Measures on Fuzzy Rough Sets. International Journal of
Pure and Applied Mathematics, 47(2008), 89-100.
[18] Shannon C.E. A Mathematical Theory of Communication. Bell System Technical Journal, 27(1948),
379-423.
[19] Sharma O., Rani A. and Gupta P. Some Similarity and Distance Measures on Fuzzy Rough Sets and
its Applications. International Journal of Engineering Sciences and Mathematics,6(2017), 85-105.
[20] Sharma O., Tiwari P. and Gupta P. (Communicated). Information Measures for Fuzzy Rough Sets and
their Applications in Pattern Recognition and Medical Area.

17
[21] Sharma O. and Gupta P. Some Trigonometric Entropies for Fuzzy Rough Set and their Application in
Medical Area. International Journal for Research in Engineering Application and Management,
4(2018), 570-578.
[22] Sharma O. Logarithmic Entropy Measures for Fuzzy Rough Set and their Application in Decision
Making. International Journal of Fuzzy System and Applications, 9(2020), 80-97.
[23] Skowron A. and Grzym.la- Busse J.W. From Rough Set Theory to Evidence Theory. In Advance in
the Dempster-Shafter Theory of Evidence, R.R. Yaeger, M. Fedrizzi and J. Kacprzyk. Eds. Wiley
New York, (1994), 193-236.
[24] Sun S. Li. L. and Hu. K. A New Approach to Rough Set Based on Remote Neighbourhood System.
Mathematical Problems in Engineering, (2019), 1-8. DOI: https://doi.org/10.1155/2019/8712010.
[25] Verbiest N. Fuzzy Rough and Evolutionary Approaches to Instance Selection. PhD Thesis, Faculty of
Sciences, Ghent University, March 2014.
[26] Zadeh L.A. Fuzzy Sets. Information and Control, 8(1965), 338-356.
[27] Zadeh L.A. Fuzzy Sets and Systems. In Proceeding of the Symposium on System Theory, (1965), 29-
37. Polytechnic Institute of Brooklyn, Brooklyn, NY, USA.
[28] Zhan J. Sun B. and Alcantud J.C. Covering Based Multi-Granulation (I.T)-Fuzzy Rough Set Models
and Application in Multi-Attribute Group Decision Making. Information Sciences, 476(2019), 290-
318.

Published Paper of International Journal of Computational Science and Information Technology (IJCSITY)

More Related Content

Similar to Published Paper of International Journal of Computational Science and Information Technology (IJCSITY) (20)

More from ijcsity123 (19)

Recently uploaded (20)

Published Paper of International Journal of Computational Science and Information Technology (IJCSITY)