INFORMATICA Informatica 1822-88440868-49520868-4952 Vilnius University INFO1142 10.15388/Informatica.2017.128 Research Article Distance Measure and Correlation Coefficient for Linguistic Hesitant Fuzzy Sets and Their Application GuanJianguan_jian@csu.edu.cn1

J. Guan received her PhD in management science and engineering from Business School of Central South University in 2006. Currently, she is a professor of management at Business School, Central South University, Changsha, China. She has contributed over 30 journal articles to professional journals. Her current research interests include decision analysis, strategic management and theory of firm in China.

 ZhouDaozhoudao@csu.edu.cn12∗

D. Zhou received his BS degree in computational mathematics from School of Mathematics and Computational Science in Xiantan University. He is a lecturer in Hunan University of Technology. Currently, he is a doctoral student in management science and engineering in Central South University. His current research interests includes fuzzy mathematics and decision making.

 MengFanyongmengfanyongtjie@163.com13

F. Meng received his PhD degree in management science and engineering from Beijing Institute of Technology in 2011. Currently, he is an associate professor in Central South University. He has contributed over 80 journal articles to professional journals such as Omega, IEEE Transactions on Systems, Man and Cybernetics Systems, Information Sciences, Knowledge-Based Systems, Applied Mathematical Modelling, Applied Mathematics and Computation, Computers and Industrial Engineering. His current research interests include fuzzy mathematics, decision making, and game theory.

 School of Business, Central South University, Changsha 410083, China School of Science, Hunan University of Technology, Zhuzhou, 412007, China Collaborative Innovation Center on Forecast and Evaluation of Meteorological Disasters, Nanjing University of Information Science and Technology, Nanjing 210044, China Corresponding author. 20171120172822372681201632017 © 2017 Vilnius University2017 Open access article under the CC BY license.

Linguistic hesitant fuzzy sets (LHFSs) permit the decision maker to apply several linguistic terms with each having several membership degrees to denote his/her preference of one thing. This type of fuzzy sets can well address the qualitative and quantitative cognitions of the decision maker as well as reflect his/her hesitancy, uncertainty and inconsistency. This paper introduces a distance measure between any two LHFSs and then defines a correlation coefficient of LHFSs. Considering the application of LHFSs, the weighted distance measure and the weighted correlation coefficient of LHFSs are defined. To address the interactions between elements in a set, the Shapley weighted distance measure and the Shapley weighted correlation coefficient are presented. It is worth noting that when the elements are independent, they degenerate to the associated weighted distance measure and the weighted correlation coefficient, respectively. After that, their application to pattern recognition is studied. Furthermore, an approach to multi-attribute decision making under linguistic hesitant fuzzy environment is developed. Meanwhile, numerical examples are offered to show the concrete application of the developed procedure.

 decision making linguistic hesitant fuzzy set correlation coefficient TOPSIS method the Shapley function Introduction

According to the attribute values of alternatives, decision-making theory can be classified into two types. One type is the stochastic decision making, where the attribute values are stochastic variables; the other is the fuzzy decision making, where the attribute values are fuzzy variables. It is worth noting that fuzzy decision-making theory has some advantages to cope with uncertain information. Since Zadeh (1965) first introduced fuzzy set theory, decision making based on fuzzy sets has been successfully applied in many fields, such as recommender system (Tejeda-Lorente et al., 2014; Martínez-Cruz et al., 2015; Yager, 2003, 2004), education (Bryson and Mobolurin, 1995), medical care (James and Dolan, 2010), engineering (Chen and Weng, 2006; Lennon et al., 2013; Meng et al., 2016c), economics (Ölçer et al., 2006; Vaidogas and Sakenaite, 2011; Meng et al., 2017a, 2016d), reservoir flood control (Fu, 2008), facility location selection (Kahraman et al., 2003), new product development (NDP) project screening (Meng and Chen, 2017a), and supplier selection (Meng et al., 2017b). With the increasing complexity of the decision-making problems, researchers found that it is insufficient to address decision-making problems by using fuzzy sets, which only permit the decision maker to apply one fuzzy number to denote the uncertainty. Furthermore, fuzzy sets can only express the decision maker’s positive judgment. Thus, several types of generalized fuzzy sets are proposed, such as intuitionistic fuzzy sets (Atanassov, 1986; Atanassov and Gargov, 1989), type-2 fuzzy sets (Zadeh, 1973) and hesitant fuzzy sets (Chen et al., 2013a; Torra, 2010).

However, all these types of fuzzy sets can only denote the decision maker’s quantitative cognitions. As Zadeh (1975) noted, there are many situations, where the decision-making problems are too complex or too ill-defined to use quantitative expressions. To address this issue, Zadeh (1975) introduced the concept of linguistic variables, which permit the decision maker to use linguistic variables rather than quantitative fuzzy variables to express the judgment. Since then, many studies about decision making based on linguistic variables are developed (Cai et al., 2014a, 2014b, 2015; Dong et al., 2009, 2016; Gou and Xu, 2016; Herrera and Martínez, 2000; Herrera et al., 2000; Ju et al., 2016; Li et al., 2017; Massanet et al., 2014; Morente-Molinera et al., 2015; Martínez and Herrera, 2012; Meng et al., 2016a; Meng and Chen, 2016b; Pedrycz, 2013; Wei, 2011; Wu and Xu, 2016; Xu, 2004a, 2007; Ye, 2016a). Just as quantitative fuzzy variables, it is still not an easy thing to require a decision maker to apply one linguistic variable to express his/her qualitative judgment. Thus, Xu (2004b) introduced the concept of uncertain linguistic variables, which permit the decision maker to use an interval linguistic variable rather than one exact linguistic variable to denote information. However, uncertain linguistic variables are inadequate to denote the decision maker’s hesitancy and irresolution. Hesitant fuzzy linguistic term sets (HFLTSs) introduced by Rodríguez et al. (2012) can well address this issue, which are composed by several linguistic terms.

All of the above mentioned fuzzy sets can denote either the decision maker’s quantitative or qualitative information. However, none of them can denote these two aspects simultaneously. Following the works of Atanassov (1986) and Zadeh (1975), Wang and Li (2009) presented intuitionistic linguistic sets (ILSs), which are composed by one linguistic variable and an intuitionistic fuzzy variable. Using this type of fuzzy sets, the decision maker can apply one linguistic variable to denote his/her qualitative judgment as well as use an intuitionistic fuzzy variable to show the membership and non-membership degrees about the qualitative judgment. Meng et al. (2016e) developed a group decision-making method with intuitionistic linguistic preference relations (ILPRs), where the elements in ILPRs are intuitionistic linguistic fuzzy variables. Later, Liu and Jin (2012) and Liu (2013) introduced intuitionistic uncertain linguistic sets (IULSs) and interval-valued intuitionistic uncertain linguistic sets (IVIULSs), respectively. Such generalizations further endow the decision makers with more rights to express their judgments. As researchers (Rodríguez et al., 2012; Torra, 2010; Ye, 2015; Meng and An, 2017) noted that the difficulty of expressing the judgments does not arise: there is a margin of error or some possibility distribution on the possibility values but there are several possible values. Recently, Meng et al. (2014) presented a new type of fuzzy sets called linguistic hesitant fuzzy sets (LHFSs). This kind of fuzzy sets permits the decision maker to apply several linguistic variables with each having several membership degrees to denote the judgment of one thing. Meanwhile, this type of fuzzy sets can express the qualitative and quantitative cognitions of the decision makers and reflect their hesitancy and inconsistency.

Considering the application of LHFSs, Meng et al. (2014) defined several operational laws of LHFSs and then gave a ranking method. After that, the authors developed a method to linguistic hesitant fuzzy multi-attribute decision making with interactive characteristics and incomplete weight information. However, one can see that this method is based on the defined aggregation operators, this makes the process of decision making seem to be complex. Especially, the calculation of the comprehensive attribute values will be very complex with the increase of the number of linguistic hesitant fuzzy sets. Later, Zhou et al. (2015) applied a special example to show that the ranking order offered in Meng et al. (2014) is unreasonable, and introduced a new ranking method. However, Zhou et al.’s ranking method is illogical. Furthermore, the Hamming distance on LHFSs offered by Zhou et al. (2015) is wrong. Recently, Zhu et al. (2016) developed a cloud model method to linguistic hesitant fuzzy multi-attribute decision making and extended the power operators to linguistic hesitant fuzzy environment. However, this method seems also to be complex.

To address the above listed issues in previous researches, the paper continues to research the application of LHFSs. To do this, we first introduce a distance measure between LHFSs that can be seen as an extension of Hamming distance on real numbers. One can check that the new distance measure addresses the issues in Zhou et al. (2015). To discriminate the importance of features or attributes, several additive weighted distance measures are defined that are used to calculate the comprehensive ranking values of objects. Meanwhile, a correlation coefficient on LHFSs is provided, and several weighted correlation coefficients are offered. Considering the interactive characteristics and the complexity of fuzzy numbers, Shapley-based distance measures and correlation coefficients with 2-additive measures are provided, which can be seen as extensions of weighted distance measures and correlation coefficients, respectively. Then, an approach to pattern recognition and to multi-attribute decision making with LHFSs is performed, respectively. Meanwhile, associated practical application is offered.

This paper is organized as follows: Section 2 reviews several basic concepts related to LHFSs. Section 3 introduces a distance measure and a correlation coefficient of LHFSs. Section 4 defines two types of hybrid weighted distance measures and correlation coefficients of LHFSs. One is based on additive measures, and the other uses the Shapley function with respect to 2-additive measures. Section 5 develops an approach to pattern recognition and to multi-attribute decision making by using the defined distance measures and correlation coefficients, and then comparison analysis with the existing methods is made. The last section is the conclusion.

 Some Basic Concepts

Considering the hesitancy and inconsistency of the decision makers, Torra (2010) defined hesitant fuzzy sets that permit the decision makers to apply several possible values in [ 0 , 1 ] to denote the membership degree of one thing.

(<italic>See</italic> Torra, <xref ref-type="bibr" rid="j_info1142_ref_061">2010</xref>).

Let X = { x 1 , x 2 , … , x n } be a finite set. A hesitant fuzzy set (HFS) in X is expressed in terms of a function such that when applied to X it returns a subset of [ 0 , 1 ], denoted by E = ( ⟨ x i , h E ( x i ) ⟩ | x i ∈ X ), where h E ( x i ) is a set of some values in [ 0 , 1 ] denoting the possible membership degrees of the element x i ∈ X to the set E.

Sometimes, it is not easy for the decision makers to estimate their information using quantitative values. In this case, linguistic variables are more suitable to only provide the decision makers with the qualitative values. The linguistic reasoning is a technique that represents qualitative aspect using linguistic variables. Let S = { s i ∣ i = 1 , 2 , … , t } be a linguistic term set with odd cardinality. Any label s i represents a possible value for a linguistic variable, and it should satisfy the following characteristics: (i) The set is ordered: s i > s j {s_{j}}$]]>, if i > jj$]]>; (ii) Max operator: max ( s i , s j ) = s i , if s i ⩾ s j ; (iii) Min operator: min ( s i , s j ) = s i , if s i ⩽ s j . For example, a linguistic term set S may be expressed by S = { s 1 : very poor, s 2 : poor, s 3 : slightly poor, s 4 : fair, s 5 : slightly good, s 6 : good, s 7 : very good}.

Similar to hesitant fuzzy sets, Rodríguez et al. (2012) introduced the following concept of hesitant fuzzy linguistic term sets (HFLTSs) that permit a qualitative reference to have several linguistic terms.

(<italic>See</italic> Rodríguez <italic>et al.</italic>, <xref ref-type="bibr" rid="j_info1142_ref_054">2012</xref>).

An HFLTS, H S , is an ordered finite subset of consecutive linguistic terms of S, where S = { s 1 , … , s t } is a linguistic term set.

For example, let S be a linguistic term set as shown above, and let Q be a qualitative reference. An HFLTS could be H S ( Q ) = { s 2 , s 3 , s 4 }.

As pointed out in introduction, HFLTSs only denote the hesitancy and inconsistency of the decision makers’ qualitative references, and it is based on the assumption that the decision makers have the same cognition degrees of the given linguistic terms in an HFLTS. However, this might be not true. For instance, to evaluate the quietness of the refrigerator, the decision maker might hesitate to give the value 15% or 20% for slightly good, the value 30%, 40% or 50% for good, and the value 15% for very good. To address this situation, HFLTSs and HFSs seem to be insufficient. Linguistic hesitant fuzzy sets (LHFSs) introduced in Meng et al. (2014) can well address this problem.

(<italic>See</italic> Meng <italic>et al.</italic>, <xref ref-type="bibr" rid="j_info1142_ref_039">2014</xref>).

Let S = { s 1 , … , s t } be a linguistic term set. A LHFS in S is a set that when applied to the linguistic terms of S it returns a subset of S with several values in [ 0 , 1 ], denoted by LH = { ( s θ ( i ) , l h ( s θ ( i ) ) ∣ s θ ( i ) ∈ S ) }, where l h ( s θ ( i ) ) = { r 1 , r 2 , … , r m i } is a set with m i values in [ 0 , 1 ] denoting the possible membership degrees of the element s θ ( i ) ∈ S to the set LH.

In the example of evaluating the quietness of the refrigerator, the decision maker’s judgment can be expressed by a LHFS LH = { ( s 5 , 0.15 , 0.2 ) , ( s 6 , 0.3 , 0.4 , 0.5 ) , ( s 7 , 0.15 ) }. To compare LHFSs, Meng et al. (2014) introduced the following method:

(<italic>See</italic> Meng <italic>et al.</italic>, <xref ref-type="bibr" rid="j_info1142_ref_039">2014</xref>).

Let LH 1 be a LHFS for the predefined linguistic term set S = { s 1 , … , s t }. Suppose that l h i = ( s θ ( i ) , l h ( s θ ( i ) ) ) ∈ LH 1 with l h ( s θ ( i ) ) = { r i 1 , r i 2 , … , r i m }, then the expectation value of l h i is defined by E ( l h i ) = θ ( i ) ∑ k = 1 m r i k m , and its variance is given as V ( l h i ) = ∑ k = 1 m ( θ ( i ) r i k − E ( l h i ) ) 2 m .

Let LH 1 be a LHFS for the predefined linguistic term set S = { s 1 , … , s t }. Suppose that l h i = ( s θ ( i ) , l h ( s θ ( i ) ) ) ∈ LH 1 with l h ( s θ ( i ) ) = { r i 1 , r i 2 , … , r i m } and l h j = ( s θ ( j ) , l h ( s θ ( j ) ) ) ∈ LH 1 with l h ( s θ ( j ) ) = { r j 1 , r j 2 , … , r j n }. Then, their order relationship is defined as follows:

If E ( l h i ) ⩽ E ( l h j ), then l h i ⩽ l h j ;

If E ( l h i ) = E ( l h i ), then V ( l h i ) > V ( l h j ) , l h i < l h j , V ( l h i ) < V ( l h j ) , l h i > l h j , V ( l h i ) = V ( l h j ) , l h i = l h j . V(l{h_{j}}),\hspace{1em}& l{h_{i}}l{h_{j}},\\ {} V(l{h_{i}})=V(l{h_{j}}),\hspace{1em}& l{h_{i}}=l{h_{j}}.\end{array}\right.$]]>

 Distance Measure and Correlation Coefficient of LHFSs

Distance measure and correlation coefficient are two useful tools to decision making, by which we can obtain the best choice or rank the objects. This section introduces a distance measure and a correlation coefficient of LHFSs.

A Distance Measure of LHFSs

Distance measure is an effective tool to measure the deviations of different arguments, which is applied in many fields, such as ranking fuzzy numbers (Tran and Duckstein, 2002), determining the weights (Yue, 2011), decision making (Cabrerizo et al., 2015; Gong et al., 2016; Peng et al., 2013; Xu, 2010b), economics (Merigó and Casanovas, 2011), pattern recognition (Hung and Yang, 2004; Zeng et al., 2016), and cluster analysis (Yang and Lin, 2009). Similar to the OWA operator (Yager, 1988), Xu and Chen (2008) defined the ordered weighted distance measure (OWDM), which can decrease the influence of extreme values. Later, Zeng and Su (2011) introduced an OWDM on intuitionistic fuzzy sets. However, the OWDM only considers the importance of the ordered positions, but it does not give the importance of the elements. It is worth noting that a distance measure corresponds to a similarity measure (Hung and Yang, 2004; Xu and Xia, 2011; Yang and Lin, 2009). This subsection gives a distance measure of LHFSs.

Let S = { s 1 , … , s t } be a linguistic term set, and let LH 1 and LH 2 be any two LHFSs for the predefined linguistic term set S. Without loss of generality, suppose that l h i = ( s θ ( i ) , l h ( s θ ( i ) ) ) ∈ LH 1 with l h ( s θ ( i ) ) = { r i 1 , r i 2 , … , r i m } and l h j = ( s θ ( j ) , l h ( s θ ( j ) ) ) ∈ LH 2 with l h ( s θ ( j ) ) = { r j 1 , r j 2 , … , r j n }. The distance measure from l h i to l h j is defined as follows: d ( l h i , l h j ) → = 1 m t ∑ k = 1 m ( min r j p ∈ l h ( s θ ( j ) ) | θ ( i ) r i k − θ ( j ) r j p | ) , and the distance measure from l h j to l h i is defined as follows: d ( l h j , l h i ) → = 1 n t ∑ k = 1 n ( min r i k ∈ l h ( s θ ( i ) ) | θ ( j ) r j k − θ ( i ) r i p | ) .

Furthermore, the distance measure between l h i and l h j is defined as follows: d ( l h i , l h j ) = d ( l h i , l h j ) → + d ( l h j , l h i ) → 2 .

Let LH 1 and LH 2 be any two LHFSs for the predefined linguistic term set S = { s 1 , … , s t }. Suppose that l h i = ( s θ ( i ) , l h ( s θ ( i ) ) ) ∈ LH 1 and l h j = ( s θ ( j ) , l h ( s θ ( j ) ) ) ∈ LH 2 are given as shown in Definition 5. Then, we have:

d ( l h i , l h j ) = 0 if and only if there is r j p ∈ l h ( s θ ( j ) ) such that θ ( j ) r j p = θ ( i ) r i k for all k = 1 , 2 , … , m, and there is r i k ∈ l h ( s θ ( i ) ) such that θ ( i ) r i k = θ ( j ) r j p for all p = 1 , 2 , … , n;

0 ⩽ d ( l h i , l h j ) ⩽ 1;

d ( l h i , l h j ) = d ( l h j , l h i );

Let LH 3 be another LHFS with l h g = ( s θ ( g ) , l h ( s θ ( g ) ) ) ∈ LH 3 . If we have min r i z ∈ l h ( s θ ( g ) ) | θ ( i ) r i k − θ ( g ) r i z | ⩽ min r j p ∈ l h ( s θ ( j ) ) | θ ( i ) r i k − θ ( j ) r j p | , min r i z ∈ l h ( s θ ( g ) ) | θ ( g ) r i z − θ ( i ) r i k | ⩽ min r i k ∈ l h ( s θ ( i ) ) | θ ( j ) r j p − θ ( i ) r i k | for all r j p ∈ l h ( s θ ( j ) ), r i k ∈ l h ( s θ ( i ) ) and r i z ∈ l h ( s θ ( g ) ), then d ( l h i , l h g ) ⩽ d ( l h i , l h j ).

For (i): when we have d ( l h i , l h j ) = 0, by the equation (3) we have d ( l h i , l h j ) → = d ( l h j , l h i ) → = 0. According to the equations (1) and (2), we get min r j p ∈ l h ( s θ ( j ) ) | θ ( i ) r i k − θ ( j ) r j p | = min r i p ∈ l h ( s θ ( i ) ) | θ ( j ) r j k − θ ( i ) r i p | = 0. Thus, there is r j p ∈ l h ( s θ ( j ) ) such that θ ( j ) r j p = θ ( i ) r i k for all k = 1 , 2 , … , m, and there is r i k ∈ l h ( s θ ( i ) ) such that θ ( i ) r i k = θ ( j ) r j p for all p = 1 , 2 , … , n. On the other hand, one can easily derive that d ( l h i , l h j ) = 0 from the equations (1)–(3) according to the listed conditions in (i).

For (ii): from the equations (1) and (3), we have 0 ⩽ d ( l h i , l h j ) → , d ( l h j , l h i ) → ⩽ 1 by 0 ⩽ min r i k ∈ l h ( s θ ( i ) ) | θ ( j ) r j k − θ ( i ) r i p |, min r j p ∈ l h ( s θ ( j ) ) | θ ( i ) r i k − θ ( j ) r j p | ⩽ t for each k = 1 , 2 , … , m and each p = 1 , 2 , … , n. According to the equation (3), we get 0 ⩽ d ( l h i , l h j ) ⩽ 1.

For (iii): from the equation (3), we obtain d ( l h i , l h j ) = d ( l h i , l h j ) → + d ( l h j , l h i ) → 2 = d ( l h j , l h i ) → + d ( l h i , l h j ) → 2 = d ( l h j , l h i ).

For (iv): from min r i z ∈ l h ( s θ ( g ) ) | θ ( i ) r i k − θ ( g ) r i z | ⩽ min r j p ∈ l h ( s θ ( j ) ) | θ ( i ) r i k − θ ( j ) r j p | min r i z ∈ l h ( s θ ( g ) ) | θ ( g ) r i z − θ ( i ) r i k | ⩽ min r i k ∈ l h ( s θ ( i ) ) | θ ( j ) r j p − θ ( i ) r i k | for all r j p ∈ l h ( s θ ( j ) ), r i k ∈ l h ( s θ ( i ) ) and r i z ∈ l h ( s θ ( g ) ), we have d ( l h i , l h g ) → ⩽ d ( l h i , l h j ) → , d ( l h g , l h i ) → ⩽ d ( l h j , l h i ) → . From the equation (3), we derive d ( l h i , l h g ) ⩽ d ( l h i , l h j ).  □

Let S = { s 1 , … , s t } be a linguistic term set, and let LH 1 and LH 2 be any two LHFSs for the predefined linguistic term set S. Without loss of generality, suppose that l h i = ( s θ ( i ) , l h ( s θ ( i ) ) ) ∈ LH 1 and l h j = ( s θ ( j ) , l h ( s θ ( j ) ) ) ∈ LH 2 . Then, the distance measure from l h i to LH 2 is defined as follows: d ( l h i , LH 2 ) → = min l h j ∈ LH 2 d ( l h i , l h j ) → , and the distance measure from l h j to LH 1 is defined as follows: d ( l h j , LH 1 ) → = min l h i ∈ LH 1 d ( l h j , l h i ) → .

Let S = { s 1 , … , s t } be a linguistic term set, and let LH 1 and LH 2 be any two LHFSs for the predefined linguistic term set S. Let l h i = ( s θ ( i ) , l h ( s θ ( i ) ) ) ∈ LH 1 , if we have d ( l h i , LH 2 ) → = d ( l h i , l h j ) with l h j ∈ LH 2 , then we denote l h j as l h j i = ( s θ i ( j ) , l h ( s θ i ( j ) ) ).

Let S = { s 1 , … , s t } be a linguistic term set, and let LH 1 and LH 2 be any two LHFSs for the predefined linguistic term set S. Then, the distance measure between LH 1 and LH 2 is defined as follows: d ( LH 1 , LH 2 ) = d ( LH 1 , LH 2 ) → + d ( LH 2 , LH 1 ) → 2 , where d ( LH 1 , LH 2 ) → = 1 | LH 1 | ∑ i = 1 | LH 1 | d ( l h i , LH 2 ) → and d ( LH 2 , LH 1 ) → = 1 | LH 2 | ∑ j = 1 | LH 2 | d ( l h j , LH 1 ) → with l h i ∈ LH 1 and l h j ∈ LH 2 , | LH 1 | and | LH 2 | denote the cardinalities of the linguistic variables in LH 1 and LH 2 , respectively.

Let LH 1 and LH 2 be any two LHFSs for the predefined linguistic term set S = { s 1 , … , s t }. Then, the distance measure d ( LH 1 , LH 2 ) between LH 1 and LH 2 satisfies:

d ( LH 1 , LH 2 ) = 0 if and only if there is l h j ∈ LH 2 such that d ( l h i , l h j ) → = 0 for any l h i ∈ LH 1 , and there is l h i ∈ LH 1 such that d ( l h j , l h i ) → = 0 for any l h j ∈ LH 2 ;

0 ⩽ d ( LH 1 , LH 2 ) ⩽ 1;

d ( LH 1 , LH 2 ) = d ( LH 2 , LH 1 );

Let LH 3 be another LHFS with l h g = ( s θ ( g ) , l h ( s θ ( g ) ) ) ∈ LH 3 . If we have min l h g ∈ LH 3 d ( l h i , l h g ) → ⩽ min l h j ∈ LH 2 d ( l h i , l h j ) → min l h i ∈ LH d ( l h g , l h i ) → ⩽ min l h i ∈ LH 1 d ( l h j , l h i ) → for all l h i ∈ LH 1 , l h j ∈ LH 2 and l h g ∈ LH 3 , then d ( LH 1 , LH 3 ) ⩽ d ( LH 1 , LH 2 ).

Similar to Property 1, one can easily derive the conclusions.  □

For example, let LH 1 = ( s 2 , 0.2 , 0.3 ) , ( s 3 , 0.3 , 0.5 , 0.7 ) , ( s 4 , 0.1 ) and LH 2 = ( s 4 , 0.5 , 0.6 ) , ( s 5 , 0.1 , 0.2 ) be two LHFSs for the predefined linguistic term set S = { s 1 , s 2 , s 3 , s 4 , s 5 , s 6 , s 7 }. Then, the distance measure between l h 1 1 = ( s 2 , 0.2 , 0.3 ) and l h 1 2 = ( s 4 , 0.5 , 0.6 ) is d ( l h 1 1 , l h 1 2 ) = d ( l h 1 1 , l h 1 2 ) → + d ( l h 1 2 , l h 1 1 ) → 2 = 1 2 × ( 1 7 × 2 ( | 2 × 0.2 − 4 × 0.5 | + | 2 × 0.3 − 4 × 0.5 | ) + 1 7 × 2 ( | 4 × 0.5 − 2 × 0.3 | + | 2 × 0.3 − 4 × 0.6 | ) ) = 0.22 , and the distance between l h 1 1 = ( s 2 , 0.2 , 0.3 ) and LH 2 is d ( l h 1 1 , LH 2 ) = min { d ( l h 1 1 , l h 1 2 ) , d ( l h 1 1 , l h 2 2 ) } = min { 0.22 , 0.025 } = 0.025.

Furthermore, the distance between LH 1 and LH 2 is d ( LH 1 , LH 2 ) = 0.034.

Let LH 1 and LH 2 be any two LHFSs for the predefined linguistic term set S = { s 1 , … , s t }. Let S ( LH 1 , LH 2 ) = 1 − d ( LH 1 , LH 2 ) . Then, S ( LH 1 , LH 2 ) is a similarity measure between LH 1 and LH 2 , which satisfies: (i) S ( LH 1 , LH 2 ) = 1 if and only if d ( LH 1 , LH 2 ) = 0; (ii) 0 ⩽ S ( LH 1 , LH 2 ) ⩽ 1; (iii) S ( LH 1 , LH 2 ) = S ( LH 2 , LH 1 ); (iv) Let LH 3 be another LHFS. If d ( LH 1 , LH 2 ) ⩽ d ( LH 1 , LH 3 ), then S ( LH 1 , LH 2 ) ⩾ S ( LH 1 , LH 3 ).

 A Correlation Coefficient of LHFSs

Correlation coefficient is a powerful tool to measure the linear relation between stochastic variables. Recently, researchers applied the correlation coefficient to measure the similarity between fuzzy variables and discussed their application in digital image processing (Van der Weken et al., 2004), clustering analysis (Chen et al., 2013b; Xu et al., 2008), pattern recognition (Liang and Shi, 2003), artificial intelligence (Zhang et al., 2013, 2014), and multi-attribute decision making (Park et al., 2009; Wei et al., 2011; Ye, 2016b; Tong and Yu, 2016). Recently, Meng and Chen (2015) noted the issues of the correlation coefficient of hesitant fuzzy sets in Chen et al. (2013b) and defined several new ones, which need not consider the lengths of HFEs and the arrangement of the possible values. Furthermore, Meng et al. (2016b) defined several correlation coefficients of interval-valued hesitant fuzzy sets in a similar way to Meng and Chen (2015). Following the work of Meng and Chen (2015), this section introduces a correlation coefficient of LHFSs.

Let S = { s 1 , … , s t } be a linguistic term set, and let LH 1 and LH 2 be any two LHFSs for the predefined linguistic term set S. The correlation coefficient between LH 1 and LH 2 is defined as follows: C C ( LH 1 , LH 2 ) = C ( LH 1 , LH 2 ) → + C ( LH 2 , LH 1 ) → max { D ( LH 1 ) , D ( LH 2 LH 1 ) } + max { D ( LH 2 ) , D ( LH 1 LH 2 ) } , where C ( LH 1 , LH 2 ) → is the correlation of LH 1 with respect to LH 2 defined as follows: C ( LH 1 , LH 2 ) → = 1 | LH 1 | ∑ l h i = ( s θ ( i ) , l h ( s θ ( i ) ) ) ∈ LH 1 1 | l h ( s θ ( i ) ) | ∑ r i k ∈ l h ( s θ ( i ) ) : r j p i k ∈ l h ( s θ i ( j ) ) , l h j i = ( s θ i ( j ) , l h ( s θ i ( j ) ) ) ∈ LH 2 θ ( i ) r i k θ i ( j ) r j p i k , with | θ ( i ) r i k − θ i ( j ) r j p i k | = min r j p i ∈ l h ( s θ i ( j ) ) | θ ( i ) r i k − θ i ( j ) r j p i | and | l h ( s θ ( i ) ) | being the cardinality of l h ( s θ ( i ) ). C ( LH 2 , LH 1 ) → is the correlation of LH 2 with respect to LH 1 defined as follows: C ( LH 2 , LH 1 ) → = 1 | LH 2 | ∑ l h j = ( s θ ( j ) , l h ( s θ ( j ) ) ) ∈ LH 2 1 | l h ( s θ ( j ) ) | ∑ r j p ∈ l h ( s θ ( j ) ) : r i k j p ∈ h ( s θ j ( i ) ) , l h i j = ( s θ j ( i ) , l h ( s θ j ( i ) ) ) ∈ LH 1 θ ( j ) r j p θ j ( i ) r i k j p , with | θ ( j ) r j p − θ j ( i ) r i k j p | = min r i k j ∈ l h ( s θ j ( i ) ) | θ ( j ) r j p − θ j ( i ) r i k j | and | l h ( s θ ( j ) ) | being the cardinality of l h ( s θ ( j ) ).

Furthermore, D ( LH 1 ) = 1 | LH 1 | ∑ l h i = ( s θ ( i ) , l h ( s θ ( i ) ) ) ∈ LH 1 1 | l h ( s θ ( i ) ) | ∑ r i k ∈ l h ( s θ ( i ) ) θ ( i ) r i k 2 , D ( LH 2 LH 1 ) = 1 | LH 1 | ∑ l h j i = ( s θ i ( j ) , l h ( s θ i ( j ) ) ) ∈ LH 2 : l h i = ( s θ ( i ) , l h ( s θ ( i ) ) ) ∈ LH 1 1 | l h ( s θ ( i ) ) | ∑ r j p i k ∈ l h ( s θ i ( j ) ) : r i k ∈ l h ( s θ ( i ) ) ( θ i ( j ) r j p i k ) 2 , D ( LH 2 ) = 1 | LH 2 | ∑ l h j = ( s θ ( j ) , l h ( s θ ( j ) ) ) ∈ LH 2 1 | l h ( s θ ( j ) ) | ∑ r j p ∈ l h ( s θ ( j ) ) ( θ ( j ) r j p ) 2 , D ( LH 1 LH 2 ) = 1 | LH 2 | ∑ l h i j = ( s θ j ( i ) , l h ( s θ j ( i ) ) ) ∈ LH 1 : l h j = ( s θ ( j ) , l h ( s θ ( j ) ) ) ∈ LH 2 1 | l h ( s θ ( j ) ) | ∑ r i k j p ∈ l h ( s θ j ( i ) ) : r j p ∈ l h ( s θ ( j ) ) ( θ j ( i ) r i k j p ) 2 .

Let S = { s 1 , … , s t } be a linguistic term set, and let LH 1 and LH 2 be any two LHFSs for the predefined linguistic term set S. The correlation coefficient between LH 1 and LH 2 satisfies:

C C ( LH 1 , LH 1 ) = 1;

C C ( LH 1 , LH 2 ) = C C ( LH 2 , LH 1 );

0 ⩽ C C ( LH 1 , LH 2 ) ⩽ 1.

For (i): From the equation (11), we have the following: D ( LH 1 LH 1 ) = 1 | LH 1 | ∑ l h j i = ( s θ i ( j ) , l h ( s θ i ( j ) ) ) ∈ LH 1 : l h i = ( s θ ( i ) , l h ( s θ ( i ) ) ) ∈ LH 1 1 | l h ( s θ ( i ) ) | ∑ r j p i k ∈ l h ( s θ i ( j ) ) : r i k ∈ l h ( s θ ( i ) ) ( θ i ( j ) r j p i k ) 2 = 1 | LH 1 | ∑ l h i = ( s θ ( i ) , l h ( s θ ( i ) ) ) ∈ LH 1 1 | l h ( s θ ( i ) ) | ∑ r i k ∈ l h ( s θ ( i ) ) ( θ ( i ) r i k ) 2 = D ( LH 1 ) .

On the other hand, from d ( l h i , LH 1 ) → = d ( l h i , l h i ) = 0, we derive the following: l h j i = ( s θ i ( j ) , l h ( s θ i ( j ) ) ) = ( s θ ( i ) , l h ( s θ ( i ) ) ) = l h i , namely, | θ ( i ) r i k − θ i ( j ) r j p i k | = min r j p i ∈ l h ( s θ i ( j ) ) | θ ( i ) r i k − θ i ( j ) r j p i | = min r i l ∈ l h ( s θ ( i ) ) | θ ( i ) r i k − θ ( i ) r i l | = | θ ( i ) r i k − θ ( i ) r i k | = 0. Thus, θ i ( j ) r j p i k = θ ( i ) r i k .

According to the equation (8), we obtain the following: C ( LH 1 , LH 1 ) → = 1 | LH 1 | ∑ l h i = ( s θ ( i ) , l h ( s θ ( i ) ) ) ∈ LH 1 1 | l h ( s θ ( i ) ) | ∑ r i k ∈ l h ( s θ ( i ) ) ( θ ( i ) r i k ) 2 = D ( LH 1 ) . Thus, C C ( LH 1 , LH 1 ) = D ( LH 1 ) D ( LH 1 ) = 1.

For (ii): From the equation (8), one can easily check that this conclusion holds.

For (iii): From C C ( LH 1 , LH 2 ) ⩾ 0, we only need to show C C ( LH 1 , LH 2 ) ⩽ 1. By the Cauchy–Schwarz inequality, we have ( C ( LH 1 , LH 2 ) → ) 2 = ( 1 | LH 1 | ∑ l h i = ( s θ ( i ) , l h ( s θ ( i ) ) ) ∈ LH 1 1 | l h ( s θ ( i ) ) | ∑ r i k ∈ l h ( s θ ( i ) ) : r j p i k ∈ l h ( s θ i ( j ) ) , ( s θ i ( j ) , l h ( s θ i ( j ) ) ) ∈ LH 2 θ ( i ) r i k θ i ( j ) r j p i k ) 2 = ( ∑ l h i = ( s θ ( i ) , l h ( s θ ( i ) ) ) ∈ LH 1 ∑ r i k ∈ l h ( s θ ( i ) ) : r j p i k ∈ l h ( s θ i ( j ) ) , ( s θ i ( j ) , l h ( s θ i ( j ) ) ) ∈ LH 2 θ ( i ) r i k θ i ( j ) r j p i k | LH 1 | | l h ( s θ ( i ) ) | ) 2 = ( ∑ l h i = ( s θ ( i ) , l h ( s θ ( i ) ) ) ∈ LH 1 ∑ r i k ∈ l h ( s θ ( i ) ) : r j p i k ∈ l h ( s θ i ( j ) ) , ( s θ i ( j ) , l h ( s θ i ( j ) ) ) ∈ LH 2 θ ( i ) r i k | LH 1 | | l h ( s θ ( i ) ) | θ i ( j ) r j p i k | LH 1 | | l h ( s θ ( i ) ) | ) 2 ⩽ ( ∑ l h i = ( s θ ( i ) , l h ( s θ ( i ) ) ) ∈ LH 1 ∑ r i k ∈ l h ( s θ ( i ) ) : r j p i k ∈ l h ( s θ i ( j ) ) , ( s θ i ( j ) , l h ( s θ i ( j ) ) ) ∈ LH 2 ( θ ( i ) r i k ) 2 | LH 1 | | l h ( s θ ( i ) ) | ) × ( ∑ l h i = ( s θ ( i ) , l h ( s θ ( i ) ) ) ∈ LH 1 ∑ r i k ∈ l h ( s θ ( i ) ) : r j p i k ∈ l h ( s θ i ( j ) ) , ( s θ i ( j ) , l h ( s θ i ( j ) ) ) ∈ LH 2 ( θ i ( j ) r j p i k ) 2 | LH 1 | | l h ( s θ ( i ) ) | ) = ( ∑ l h i = ( s θ ( i ) , l h ( s θ ( i ) ) ) ∈ LH 1 ∑ r i k ∈ l h ( s θ ( i ) ) ( θ ( i ) r i k ) 2 | LH 1 | | l h ( s θ ( i ) ) | ) × ( ∑ l h j i = ( s θ i ( j ) , l h ( s θ i ( j ) ) ) ∈ LH 2 : l h i = ( s θ ( i ) , l h ( s θ ( i ) ) ) ∈ LH 1 ∑ r j p i k ∈ l h ( s θ i ( j ) ) : r i k ∈ l h ( s θ ( i ) ) ( θ i ( j ) r j p i k ) 2 | LH 1 | | l h ( s θ ( i ) ) | ) = D ( LH 1 ) D ( LH 2 LH 1 ) .

From D ( LH 1 ) D ( LH 2 LH 1 ) ⩽ D ( LH 1 ) + D ( LH 2 LH 1 ) 2 ⩽ max { D ( LH 1 ) , D ( LH 2 LH 1 ) }, we derive C ( LH 1 , LH 2 ) → ⩽ max { D ( LH 1 ) , D ( LH 2 LH 1 ) } . Similarly, we have C ( LH 2 , LH 1 ) → ⩽ max { D ( LH 2 ) , D ( LH 1 LH 2 ) }. Thus, C C ( LH 1 , LH 2 ) ⩽ 1 .  □

Similar to the correlation coefficient defined in the equation (7), we can define other correlation coefficients of LHFSs in a similar way to Meng and Chen (2015). When the membership degree of each linguistic term in LHFSs is one, then LHFSs degenerate to HFLTSs. In this situation, the correlation coefficient (7) reduces to the correlation coefficient for HFLTSs. Furthermore, when all LHFSs only have the same one linguistic term, then the correlation coefficient (7) reduces to the correlation coefficient for HFSs given by Meng and Chen (2015).

For example, let LH 1 and LH 2 be two LHFSs for the predefined linguistic term set S = { s 1 , s 2 , s 3 , s 4 , s 5 , s 6 , s 7 }, where LH 1 and LH 2 as shown above, namely, LH 1 = { ( s 2 , 0.2 , 0.3 ) , ( s 3 , 0.3 , 0.5 , 0.7 ) , ( s 4 , 0.1 ) } and LH 2 = { ( s 4 , 0.5 , 0.6 ) , ( s 5 , 0.1 , 0.2 ) }. Then, the correlation coefficient between LH 1 and LH 2 is C C ( LH 1 , LH 2 ) = 0.8747.

 Hybrid Weighted Distance Measure and Correlation Coefficient of LHFSs

This section studies two types of the hybrid weighted distance measures and the hybrid weighted correlation coefficients of LHFSs. One is based on additive measures, which does not consider the interactions between elements in a set; the other adopts 2-additive measures to reflect their interactive characteristics. It is worth noting that the former can be seen as a special case of the latter.

Distance Measure and Correlation Coefficient Based on Additive Measures

To both consider the importance of the elements and their ordered positions, this subsection defines the hybrid weighted distance measure and the hybrid weighted correlation coefficient based on additive measures.

Let S = { s 1 , … , s t } be a linguistic term set, and let A = { LH 1 , LH 2 , … , LH n } and B = { LH 1 ′ , LH 2 ′ , … , LH n ′ } be any two collections of LHFSs for the predefined linguistic term set S. Then, the hybrid weighted distance measure (HWDM) between A and B is defined as follows: HWDM ( A , B ) = ∑ i = 1 n w i ω ( i ) ∑ i = 1 n w i ω ( i ) d ( LH ( i ) , LH ( i ) ′ ) , and the geometric hybrid weighted distance measure (GHWDM) between A and B is defined as follows: GHWDM ( A , B ) = ∏ i = 1 n d ( LH ( i ) , LH ( i ) ′ ) w i ω ( i ) ∑ i = 1 n w i ω ( i ) , where ( · ) is a permutation on the distance measures d ( LH j , LH j ′ ), j = 1 , 2 , … , n, with ω ( j ) d ( LH ( j ) , LH ( j ) ′ ) being the jth largest value of ω j d ( LH j , LH j ′ ) and d ( LH ( j ) , LH ( j ) ′ ) ω ( j ) being the jth largest value of d ( LH j , LH j ′ ) ω j , w = { w 1 , w 2 , … , w n } is the weighting vector on the ordered position set N = { 1 , 2 , … , n } such that ∑ i = 1 n w i = 1 and w i ⩾ 0 for all i = 1 , 2 , … , n, and ω = ( ω 1 , ω 2 , … , ω n ) is the weighting vector on D = { d ( LH i , LH i ′ ) } i ∈ N such that ∑ i = 1 n ω i = 1 and ω i ⩾ 0 for all i = 1 , 2 , … , n.

Let A = { LH 1 , LH 2 , … , LH n } and B = { LH 1 ′ , LH 2 ′ , … , LH n ′ } be any two collections of LHFSs for the predefined linguistic term set S = { s 1 , … , s t }. Then, GHWDM ( A , B ) ⩽ HWDM ( A , B ).

According to ∑ i = 1 n ω i a i ⩽ ∏ i = 1 n a i ω i with a i > 00$]]> for all i = 1 , 2 , … , n, where ω is a weighting vector as shown in Definition 9, we have GHWDM ( A , B ) ⩽ HWDM ( A , B ) from their expressions.  □

Let A = { LH 1 , LH 2 , … , LH n } and B = { LH 1 ′ , LH 2 ′ , … , LH n ′ } be any two collections of LHFSs for the predefined linguistic term set S = { s 1 , … , s t }.

Commutativity. Let σ be a permutation of A and B, respectively, where A ˙ = { LH σ ( 1 ) , LH σ ( 2 ) , … , LH σ ( n ) } and B ˙ = { LH σ ( 1 ) ′ , LH σ ( 2 ) ′ , … , LH σ ( n ) ′ }. Then, HWDM ( A , B ) = HWDM ( A ˙ , B ˙ ) and GHWDM ( A , B ) = GHWDM ( A ˙ , B ˙ ) .

Monotonicity. Let C = { C a u c h y – S c h w a r z 1 ″ , C a u c h y – S c h w a r z 2 ″ , … , LH n ″ } be another collection of LHFSs for the predefined linguistic term set S, where d ( LH i , LH i ′ ) ⩽ d ( LH i , LH i ″ ) for all i = 1 , 2 , … , n. Then, HWDM ( A , B ) ⩽ HWDM ( A , C ) and GHWDM ( A , B ) ⩽ GHWDM ( A , C ) .

Idempotency. If d ( LH i , LH i ′ ) = c for all i = 1 , 2 , … , n, then HWDM ( A , B ) = c and GHWDM ( A , B ) = c .

Boundary. min 1 ⩽ i ⩽ n d ( LH i , LH i ′ ) ⩽ HWDM ( A , B ) ⩽ max 1 ⩽ i ⩽ n d ( LH i , LH i ′ ) , min 1 ⩽ i ⩽ n d ( LH i , LH i ′ ) ⩽ GHWDM ( A , B ) ⩽ max 1 ⩽ i ⩽ n d ( LH i , LH i ′ ) .

For (i): we have HWDM ( A , B ) = ∑ i = 1 n w i ω ( i ) ∑ i = 1 n w i ω ( i ) d ( LH ( i ) , LH ( i ) ′ ) = ∑ i = 1 n w σ ( i ) ω ( σ ( i ) ) ∑ i = 1 n w σ ( i ) ω ( σ ( i ) ) d ( LH ( σ ( i ) ) , LH ( σ ( i ) ) ′ ) = HWDM ( A ˙ , B ˙ ) . Similarly, we get GHWDM ( A , B ) = GHWDM ( A ˙ , B ˙ ).

For (ii): from Definition 9 and d ( LH i , LH i ′ ) ⩽ d ( LH i , L H i ″ ) for all i = 1 , 2 , … , n, one can easily derive the conclusion.

For (iii): from d ( LH i , LH i ′ ) = c for all i = 1 , 2 , … , n, we get HWDM ( A , B ) = ∑ i = 1 n w i ω ( i ) ∑ i = 1 n w i ω ( i ) d ( LH ( i ) , LH ( i ) ′ ) = ∑ i = 1 n w i ω ( i ) ∑ i = 1 n w i ω ( i ) c = c ∑ i = 1 n w i ω ( i ) ∑ i = 1 n w i ω ( i ) = c . In a similar way, we have GHWDM ( A , B ) = c.

For (iv): because min 1 ⩽ i ⩽ n d ( LH i , LH i ′ ) ⩽ d ( LH ( i ) , LH ( i ) ′ ) ⩽ max 1 ⩽ i ⩽ n d ( LH i , LH i ′ ) for all i = 1 , 2 , … , n, we derive ∑ i = 1 n w i ω ( i ) ∑ i = 1 n w i ω ( i ) min 1 ⩽ i ⩽ n d ( LH i , LH i ′ ) ⩽ ∑ i = 1 n w i ω ( i ) ∑ i = 1 n w i ω ( i ) d ( LH ( i ) , LH ( i ) ′ ) ⩽ ∑ i = 1 n w i ω ( i ) ∑ i = 1 n w i ω ( i ) max 1 ⩽ i ⩽ n d ( LH i , LH i ′ ) and min 1 ⩽ i ⩽ n d ( LH i , LH i ′ ) ⩽ ∑ i = 1 n w i ω ( i ) ∑ i = 1 n w i ω ( i ) d ( LH ( i ) , LH ( i ) ′ ) ⩽ max 1 ⩽ i ⩽ n d ( LH i , LH i ′ ) . Thus, min 1 ⩽ i ⩽ n d ( LH i , LH i ′ ) ⩽ HWDM ( A , B ) ⩽ max 1 ⩽ i ⩽ n d ( LH i , LH i ′ ) . In a similar way, one can derive min 1 ⩽ i ⩽ n d ( LH i , LH i ′ ) ⩽ GHWDM ( A , B ) ⩽ max 1 ⩽ i ⩽ n d ( LH i , LH i ′ ).  □

When w i = 1 / n for all i = 1 , 2 , … , n, then the HWDM degenerates to the weighted distance measure (WDM): WDM ( A , B ) = ∑ i = 1 n ω i d ( LH i , LH i ′ ) , and the GHWDM degenerates to the geometric weighted distance measure (GWDM): GWDM ( A , B ) = ∏ i = 1 n d ( LH i , LH i ′ ) ω i . When ω i = 1 / n for all i = 1 , 2 , … , n, then the HWDM degenerates to the ordered weighted distance measure (OWDM): OWDM ( A , B ) = ∑ i = 1 n w i d ( LH ( i ) , LH ( i ) ′ ) , and the GHWDM degenerates to the geometric ordered weighted distance measure (GOWDM): GOWDM ( A , B ) = ∏ i = 1 n d ( LH ( i ) , LH ( i ) ′ ) w i .

Let S = { s 1 , … , s t } be a linguistic term set, and let A = { LH 1 , LH 2 , … , LH n } and B = { LH 1 ′ , LH 2 ′ , … , LH n ′ } be any two collections of LHFSs for the predefined linguistic term set S. Then, the hybrid weighted correlation coefficient (HWCC) between A and B is defined as follows: HWCC ( A , B ) = ∑ i = 1 n w i ω ( i ) ∑ i = 1 n w i ω ( i ) C C ( LH ( i ) , LH ( i ) ′ ) , and the geometric hybrid weighted correlation coefficient (GHWCC) between A and B is defined as follows: GHWCC ( A , B ) = ∏ i = 1 n C C ( LH ( i ) , LH ( i ) ′ ) w i ω ( i ) ∑ i = 1 n w i ω ( i ) where ( · ) is a permutation on the correlation coefficients C C ( LH j , LH j ′ ), j = 1 , 2 , … , n, with ω ( j ) C C ( LH ( j ) , LH ( j ) ′ ) being the jth largest value of ω j C C ( LH j , LH j ′ ) and C C ( LH ( j ) , LH ( j ) ′ ) ω ( j ) being the jth largest value of C C ( LH j , LH j ′ ) ω j , w = { w 1 , w 2 , … , w n } is the weighting vector on the ordered position set N = { 1 , 2 , … , n } such that ∑ i = 1 n w i = 1 and w i ⩾ 0 for all i = 1 , 2 , … , n, and ω = ( ω 1 , ω 2 , … , ω n ) is the weighting vector on D = { d ( LH i , LH i ′ ) } i ∈ N such that ∑ i = 1 n ω i = 1 and ω i ⩾ 0 for all i = 1 , 2 , … , n.

When w i = 1 / n for all i = 1 , 2 , … , n, then the HWCC degenerates to the weighted correlation coefficient (WCC): WCC ( A , B ) = ∑ i = 1 n ω i C C ( LH i , LH i ′ ) , and the GHWCC degenerates to the geometric weighted correlation coefficient (GWCC): GWCC ( A , B ) = ∏ i = 1 n C C ( LH i , LH i ′ ) ω i . When ω i = 1 / n for all i = 1 , 2 , … , n, then the HWCC degenerates to the ordered weighted correlation coefficient (OWCC): OWCC ( A , B ) = ∑ i = 1 n w i C C ( LH ( i ) , LH ( i ) ′ ) , and the GHWCC degenerates to the geometric ordered weighted correlation coefficient (GOWCC): GOWCC ( A , B ) = ∏ i = 1 n C C ( LH ( i ) , LH ( i ) ′ ) w i .

Let A = { LH 1 , LH 2 , … , LH n } and B = { LH 1 ′ , LH 2 ′ , … , LH n ′ } be any two collections of LHFSs for the predefined linguistic term set S = { s 1 , … , s t }. Then, GHWCC ( A , B ) ⩽ HWCC ( A , B ).

Similar to the HWDM and the GHWDM, one can check that the HWCC and the GHWCC satisfy the properties listed in Property 4.

 Distance Measure and Correlation Coefficient Based on 2-Additive Measures

The hybrid weighted distance measure and correlation coefficient defined in Section 4.1 are all based on the assumption that the importance of elements in a set is independent. However, in some situations, this assumption is incorrect (Grabisch, 1997; Meng et al., 2016b, 2016f; Meng and Chen, 2017b; Meng and Chen, 2016a, 2016c; Tan, 2011; Tan and Chen, 2011; Xu, 2010a). Considering the interactions between elements, this subsection uses the Shapley function (Shapley, 1953) with respect to 2-additive measures (Grabisch, 1997) to define the hybrid 2-additive Shapley distance measure and the hybrid 2-additive Shapley correlation coefficient. Sugeno (1974) introduced the following concept of fuzzy measures:

(<italic>See</italic> Sugeno, <xref ref-type="bibr" rid="j_info1142_ref_056">1974</xref>).

A fuzzy measure on finite set N = { 1 , 2 , … , n } is a set function μ : P ( N ) → [ 0 , 1 ] satisfying:

μ ( ∅ ) = 0 , μ ( N ) = 1;

For all A , B ∈ P ( N ) with A ⊆ B, μ ( A ) ⩽ μ ( B ),

 where P ( N ) is the power set of N.

From Definition 11, we know that fuzzy measures do not only give the importance of elements separately but also consider the importance of all their combinations. For any two coalitions A , B ∈ P ( N ) with A ∩ B = ∅, when μ ( A ) + μ ( B ) < μ ( A ∪ B ), then there exists complementary interaction between A and B; when μ ( A ) + μ ( B ) > μ ( A ∪ B )\mu (A\cup B)$]]>, then there exists redundant interaction between A and B; otherwise, there is no interaction between A and B, namely, μ ( A ) + μ ( B ) = μ ( A ∪ B ). Although fuzzy measures well reflect the interactions between elements, they define on the power set. This means that it is not an easy thing to determine a fuzzy measure. To cope with this problem, 2-additive measures (Grabisch, 1997) are a good choice.

(<italic>See</italic> Grabisch, <xref ref-type="bibr" rid="j_info1142_ref_016">1997</xref>).

The fuzzy measure μ on N = { 1 , 2 , … , n } is called a 2-additive measure, if, for any S ⊆ N with s ⩾ 2 , we have μ ( S ) = ∑ { i , j } ⊆ S μ ( i , j ) − ( s − 2 ) ∑ i ∈ S μ ( i ) , where s is the cardinality of S.

According to Definition 12, when we know the importance of each element and all of their combined in pairs, then we can derive an associated 2-additive measure. Following the work of Meng and Tang (2013), we apply the Shapley function (Shapley, 1953) with respect to 2-additive measures to give the weights of the elements.

(See Meng and Tang, <xref ref-type="bibr" rid="j_info1142_ref_038">2013</xref>).

Let v be a 2-additive measure defined on N = { 1 , 2 , … , n }, then the Shapley function for v can be expressed as follows: ϕ i ( N , v ) = 3 − n 2 v ( i ) + 1 2 ∑ j ∈ N ∖ i ( v ( i , j ) − v ( j ) ) .

Now, let us define the following distance measure and correlation coefficient using the Shapley function with respect to 2-additive measures.

Let S = { s 1 , … , s t } be a linguistic term set, and let A = { LH 1 , LH 2 , … , LH n } and B = { LH 1 ′ , LH 2 ′ , … , LH n ′ } be any two collections of LHFSs for the predefined linguistic term set S. Then, the hybrid 2-additive Shapley distance measure (H2SDM) between A and B is defined as follows: H 2 SDM ( A , B ) = ∑ i = 1 n ϕ i ( N , v ) φ ( i ) ( D , μ ) ∑ i = 1 n ϕ i ( N , v ) φ ( i ) ( D , μ ) d ( LH ( i ) , LH ( i ) ′ ) , and the geometric hybrid 2-additive Shapley distance measure (GH2SDM) between A and B is defined as follows: GH 2 SDM ( A , B ) = ∏ i = 1 n d ( LH ( i ) , LH ( i ) ′ ) ϕ i ( N , v ) φ ( i ) ( D , μ ) ∑ i = 1 n ϕ i ( N , v ) φ ( i ) ( D , μ ) , where ( · ) is a permutation on the distance measures d ( LH j , LH j ′ ) , j = 1 , 2 , … , n, with φ ( j ) ( D , μ ) d ( LH ( j ) , LH ( j ) ′ ) being the jth largest value of φ j ( D , μ ) d ( LH j , LH j ′ ) and d ( LH ( j ) , LH ( j ) ′ ) φ ( j ) ( D , μ ) being the jth largest value of d ( LH j , LH j ′ ) φ j ( D , μ ) , ϕ i ( N , v ) is the Shapley value of the ith position with respect to the 2-additive measure v on the ordered position set N = { 1 , 2 , … , n }, and φ i ( D , μ ) is the Shapley value of the distance measure d ( LH i , LH i ′ ) with respect to the 2-additive measure μ on D = { d ( LH i , LH i ′ ) } i ∈ N .

When there are no interactions between elements as well as between the ordered positions, then the H2SDM reduces to the HWDM, and the GH2SDM reduces to the GHWDM. Furthermore, when ϕ i ( N , v ) = 1 / n for all i = 1 , 2 , … , n, then the H2SDM degenerates to the 2-additive Shapley distance measure (2SDM): 2 S D M ( A , B ) = ∑ i = 1 n φ i ( D , μ ) d ( LH i , LH i ′ ) , and the GH2SDM degenerates to the geometric 2-additive Shapley distance measure (G2SDM): G 2 S D M ( A , B ) = ∏ i = 1 n d ( LH i , LH i ′ ) φ i ( D , μ ) . When φ i ( D , μ ) = 1 / n for all i = 1 , 2 , … , n, then the H2SDM degenerates to the ordered 2-additive Shapley distance measure (O2SDM): O 2 S D M ( A , B ) = ∑ i = 1 n ϕ i ( N , v ) d ( LH ( i ) , LH ( i ) ′ ) , and the GH2SDM degenerates to the geometric ordered 2-additive Shapley distance measure (GO2SDM): G O 2 S D M ( A , B ) = ∏ i = 1 n d ( LH ( i ) , LH ( i ) ′ ) ϕ i ( N , v ) .

Let S = { s 1 , … , s t } be a linguistic term set, and let A = { LH 1 , LH 2 , … , LH n } and B = { LH 1 ′ , LH 2 ′ , … , LH n ′ } be any two collections of LHFSs for the predefined linguistic term set S. Then, the hybrid 2-additive Shapley correlation coefficient (H2SCC) between A and B is defined as follows: H 2 S C C ( A , B ) = ∑ i = 1 n ϕ i ( N , v ) φ ( i ) ( E , μ ) ∑ i = 1 n ϕ i ( N , v ) φ ( i ) ( E , μ ) C C ( LH ( i ) , LH ( i ) ′ ) , and the geometric hybrid 2-additive Shapley correlation coefficient (GH2SCC) between A and B is defined as follows: G H 2 S C C ( A , B ) = ∏ i = 1 n C C ( LH ( i ) , LH ( i ) ′ ) ϕ i ( N , v ) φ ( i ) ( E , μ ) ∑ i = 1 n ϕ i ( N , v ) φ ( i ) ( E , μ ) , where ( · ) is a permutation on the correlation coefficients C C ( LH j , LH j ′ ), j = 1 , 2 , … , n, with φ ( j ) ( E , μ ) C C ( LH ( j ) , LH ( j ) ′ ) being the jth largest value of φ j ( E , μ ) C C ( LH j , LH j ′ ) and C C ( LH ( j ) , LH ( j ) ′ ) φ ( j ) ( E , μ ) being the jth largest value of C C ( LH j , LH j ′ ) φ j ( E , μ ) , ϕ i ( N , v ) is the Shapley value of the ith position with respect to the 2-additive measures on the ordered position set N = { 1 , 2 , … , n }, and φ i ( E , μ ) is the Shapley value of the distance measure C C ( LH i , LH i ′ ) with respect to the 2-additive measure μ on E = { C C ( LH i , LH i ′ ) } i ∈ N .

When there are no interactions between elements as well as between the ordered positions, then the H2SCC reduces to the HWCC, and the GH2SCC reduces to the GHWCC. Furthermore, when ϕ i ( N , v ) = 1 / n for all i = 1 , 2 , … , n, then the H2SCC degenerates to the 2-additive Shapley correlation coefficient (2SCC): 2 S C C ( A , B ) = ∑ i = 1 n φ i ( E , μ ) C C ( LH i , LH i ′ ) , and the GH2SDM degenerates to the geometric 2-additive Shapley correlation coefficient (G2SCC): G 2 S C C ( A , B ) = ∏ i = 1 n C C ( LH i , LH i ′ ) φ i ( E , μ ) . When φ i ( D , μ ) = 1 / n for all i = 1 , 2 , … , n, then the H2SCC degenerates to the ordered 2-additive Shapley correlation coefficient (O2SCC): O 2 S C C ( A , B ) = ∑ i = 1 n ϕ i ( N , v ) C C ( LH ( i ) , LH ( i ) ′ ) , and the GH2SCC degenerates to the geometric ordered 2-additive Shapley correlation coefficient (GO2SCC): G O 2 S C C ( A , B ) = ∏ i = 1 n C C ( LH ( i ) , LH ( i ) ′ ) ϕ i ( N , v ) .

Furthermore, one can easily show that all distance measures and correlation coefficients defined in this subsection satisfy the properties given in Property 4.

 Application to Pattern Recognition and to Multi-Attribute Decision Making

This section researches the application of the distance measure and the correlation coefficient to pattern recognition and to multi-attribute decision making. Additionally, a comparison analysis is made to validate the effectiveness of the proposed approach.

The Application to Pattern Recognition

This subsection presents an approach to pattern recognition with LHFSs using the distance measure and the correlation coefficient. The main decision procedure can be described as follows:

Step 1:

Suppose that there are m patterns A = { A 1 , A 2 , … , A m } and n features C = { c 1 , c 2 , … , c n }. Let S = { s 1 , … , s t } be the predefined linguistic term set. The evaluation of each pattern A i with respect to the feature c j is a LHFS, denoted by A i = { ⟨ c j , LH i j = { ( s θ ( i j ) , l h ( s θ ( i j ) ) | s θ ( i j ) ∈ S ) } ⟩ | j = 1 , 2 , … , n } for each i = 1 , 2 , … , m. Furthermore, assume there is a sample ε to be recognized, which is represented by ε = { ⟨ c j , LH j = { ( s θ ( j ) , l h ( s θ ( j ) ) | s θ ( j ) ∈ S ) } ⟩ | j = 1 , 2 , … , n } .

Step 2:

Use the HWDM or GHWDM to calculate the distance measure between A i ( i = 1 , 2 , … , m ) and ε, or adopt the H2SDM or GH2SDM to calculate the distance measure between A i ( i = 1 , 2 , … , m ) and ε.

Step 3:

According to the distance measure, identify the most likely pattern.

Step 3:

End.

Similar to the distance measure, we can apply the correlation coefficient to give a method to pattern recognition. (<italic>See</italic> Zhang and Jiang, <xref ref-type="bibr" rid="j_info1142_ref_092">2008</xref>).

Let us consider a set of diagnoses A = { A 1 ( Viral fever ) , A 2 ( Malaria ) , A 3 ( Typhoid ) }, and a set of symptoms C = { c 1 : temperature; c 2 : headache; c 3 : cough}. Let S 1 = { s 1 : fair, s 2 : a little high, s 3 : high, s 4 : very high, s 5 : extremely high} be the predefined linguistic term set for the feature c 1 , and let S 2 = { s 1 : extremely slight, s 2 : very slight, s 3 : slight, s 4 : a little slight, s 5 : fair, s 6 : a little heavy, s 7 : heavy; s 8 : very heavy s 9 : extremely heavy} be the predefined linguistic term set for the features c 2 and c 3 . Suppose that a patient, with respect to all the symptoms, can be represented by the following LHFS: ε ( patient ) = { ⟨ c 1 , { ( s 4 , 0.5 , 0.7 ) , ( s 5 , 0.4 ) } ⟩ , ⟨ c 2 , { ( s 6 , 0.6 ) , ( s 7 , 0.5 , 0.8 ) } ⟩ , ⟨ c 3 , { ( s 4 , 0.6 ) } ⟩ } and assume that each diagnosis A i ( i = 1 , 2 , 3 ) is viewed as a LHFS with respect to all the symptoms, where A 1 = { ⟨ c 1 , { ( s 2 , 0.3 ) , ( s 3 , 0.6 , 0.7 ) } ⟩ , ⟨ c 2 , { ( s 5 , 0.7 , 0.9 ) , ( s 6 , 0.4 ) } ⟩ , ⟨ c 3 , { ( s 6 , 0.4 , 0.6 ) , ( s 7 , 0.7 ) } ⟩ } , A 2 = { ⟨ c 1 , { ( s 4 , 0.6 , 0.8 ) } ⟩ , ⟨ c 2 , { ( s 4 , 0.5 , 0.7 ) , ( s 5 , 0.6 , 0.8 ) } ⟩ , ⟨ c 3 , { ( s 6 , 0.5 , 0.7 ) } ⟩ } , A 3 = { ⟨ c 1 , { ( s 4 , 0.3 , 0.5 ) , ( s 5 , 0.7 ) } ⟩ , ⟨ c 2 , { ( s 6 , 0.6 , 0.8 ) } ⟩ , ⟨ c 3 , { ( s 4 , 0.3 , 0.6 ) , ( s 5 , 0.7 , 0.8 ) } ⟩ } . Our aim is to classify the patient to one of the diagnoses A 1 , A 2 and A 3 . Assume that the importance of features is given by 0.5, 0.2 and 0.3, and the weights of the ordered positions are defined by 0.3, 0.4 and 0.3.

Using the HWDM, we derive: HWDM ( A 1 , ε ) = 0.6525 , HWDM ( A 2 , ε ) = 0.5907 , HWDM ( A 3 , ε ) = 0.4166. From HWDM ( A 1 , ε ) ⩾ HWDM ( A 2 , ε ) ⩾ HWDM ( A 3 , ε ), we know that the patient suffers from typhoid.

When the GHWDM is applied to calculate the distance measure, we obtain: GHWDM ( A 1 , ε ) = 0.6286 , GHWDM ( A 2 , ε ) = 0.5904 , GHWDM ( A 3 , ε ) = 0.3932 , which also shows that the patient has typhoid.

Furthermore, when the HWCC and GHWCC are used to compute the correlation coefficients, we get: HWCC ( A 1 , ε ) = 0.7747 , HWCC ( A 2 , ε ) = 0.8169 , HWCC ( A 3 , ε ) = 0.8744 and GHWCC ( A 1 , ε ) = 0.7705 , GHWCC ( A 2 , ε ) = 0.8321 , GHWCC ( A 3 , ε ) = 0.889 which means that the patient has typhoid too.

All of the above distance measures and correlation coefficients only consider the importance of elements separately. However, the importance of their combinations is not given. Considering the following facts, the symptoms of temperature and cough give more information than that of headache and temperature to diagnose the patient, and the symptoms of headache and temperature give more information than that of headache and cough. Suppose that the importance of temperature and cough, headache and temperature, and headache and cough is respectively defined by 0.9, 0.7, and 0.4. Furthermore, the importance of the ordered positions 1 and 2 is equal to that of the ordered positions 2 and 3, which both equal 0.75. However, the importance of the ordered positions 1 and 3 is less than that of the ordered positions 1 and 2 or 2 and 3, which is defined by 0.5. In this case, the distance measures and the correlation coefficients based on additive measures are helpless. However, the distance measures and the correlation coefficients using the Shapley function with respect to 2-additive measures are good choices to deal with this situation.

From the above depiction, we know that the 2-additive measure μ on the feature set C is μ ( ∅ ) = 0 , μ ( c 1 ) = 0.5 , μ ( c 2 ) = 0.2 , μ ( c 3 ) = 0.3 , μ ( c 1 , c 2 ) = 0.9 , μ ( c 1 , c 3 ) = 0.7 , μ ( c 2 , c 3 ) = 0.4. Furthermore, the 2-additive measure v on the ordered position set N is v ( ∅ ) = 0, v ( 1 ) = 0.3, v ( 2 ) = 0.4, v ( 3 ) = 0.3, v ( 1 , 2 ) = v ( 2 , 3 ) = 0.75, v ( 1 , 3 ) = 0.5. Using the equation (15), the Shapley values of the features are ϕ c 1 ( C , μ ) = 0.55 , ϕ c 2 ( C , μ ) = 0.25 , ϕ c 3 ( C , μ ) = 0.2 and the Shapley values of the ordered positions are ϕ 1 ( N , v ) = ϕ 3 ( N , v ) = 0.275 , ϕ 2 ( N , v ) = 0.45. Using the H2SDM and the GH2SDM, we obtain: H 2 SDM ( A 1 , ε ) = 0.6196 , H 2 SDM ( A 2 , ε ) = 0.6142 , H 2 SDM ( A 3 , ε ) = 0.3993 and GH 2 SDM ( A 1 , ε ) = 0.6105 , GH 2 SDM ( A 2 , ε ) = 0.5777 , GH 2 SDM ( A 3 , ε ) = 0.3780 which both show that the patient suffers from typhoid.

Using the H2SCC and the GH2SCC, we get: H 2 S C C ( A 1 , ε ) = 0.8154 , H 2 S C C ( A 2 , ε ) = 0.8469 , H 2 S C C ( A 3 , ε ) = 0.8872 and G H 2 S C C ( A 1 , ε ) = 0.7895 , G H 2 S C C ( A 2 , ε ) = 0.8414 , G H 2 S C C ( A 3 , ε ) = 0.8655 which still show that the patient has typhoid.

It is interesting that all distance measures and correlation coefficients show the patient having typhoid. However, their ranking values are different. In practical application, when we only need to consider the importance of each element, then the decision makers can apply the distance measures and correlation coefficients based on additive measures; otherwise, we suggest that the decision makers use the 2-additive measure based distance measures and correlation coefficients.

 An Approach to Multi-Attribute Decision Making

This subsection introduces a new decision-making method with linguistic hesitant fuzzy information. Considering a multi-attribute decision making problem, let S = { s 1 , … , s t } be the predefined linguistic term set. Suppose that there are m alternatives A = { A 1 , A 2 , … , A m } to be evaluated according to n attributes C = { c 1 , c 2 , … , c n }. The main steps are given as follows:

Step 1:

Assume that the evaluation of the alternative A i with respect to the attribute c j is a LHFS LH i j = { ( s θ ( i j ) , l h ( s θ ( i j ) ) ∣ s θ ( i j ) ∈ S ) } ( i = 1 , 2 , … , m; j = 1 , 2 , … , n). Let LH = ( LH i j ) m × n be the evaluation LHFS matrix.

Step 2:

If all attributes c j ( j = 1 , 2 , … , n ) are benefits (i.e. the larger, the greater preference), then the attribute values need not to be normalized. Otherwise, we normalize the decision matrix LH = ( LH i j ) m × n into LH ′ = ( LH i j ′ ) m × n , where LH i j ′ = LH i j for benefit criterion c j ( LH i j ′ ) C for cost criterion c j ( i = 1 , 2 , … , m ; j = 1 , 2 , … , n ). ( LH i j ′ ) C is the complement of LH i j such that ( LH i j ′ ) C = { ( s t − θ ( i j ) , ⋃ r i j ∈ l h ( s θ ( i j ) ) ( 1 − r i j ) ∣ s θ ( i j ) ∈ S ) }.

Step 3:

Give the positive-ideal LHFS LH j + = ( s t , 1 ) and the negative-ideal LHFS LH j − = ( s 1 , 0 ) for each j = 1 , 2 , … , n. Let LH + = { LH 1 ′ + , LH 2 ′ + , … , LH n ′ + } and LH − = { LH 1 ′ − , LH 2 ′ − , … , LH n ′ − } .

Step 4:

Apply the HWDM or GHWDM to calculate the distance measure between LH i = { LH i 1 , LH i 2 , … , LH i n } ( i = 1 , 2 , … , m ) and LH + as well as the distance measure between LH i ( i = 1 , 2 , … , m ) and LH − , or we adopt the H2SDM or GH2SDM to calculate the distance measure between LH i ( i = 1 , 2 , … , m ) and LH + as well as the distance measure between LH i ( i = 1 , 2 , … , m ) and LH − .

Step 5:

Calculate the similarity measure D i for the alternative A i ( i = 1 , 2 , … , m ), where D i = 1 − HWDM ( LH i , M + ) HWDM ( LH i , M + ) + HWDM ( LH i , M − ) , D i = 1 − GHWDM ( LH i , M + ) GHWDM ( LH i , M + ) + GHWDM ( LH i , M − ) , D i = 1 − H 2 SDM ( LH i , M + ) H 2 SDM ( LH i , M + ) + H 2 SDM ( LH i , M − ) , D i = 1 − GH 2 SDM ( LH i , M + ) GH 2 SDM ( LH i , M + ) + GH 2 SDM ( LH i , M − ) .

According to D i ( i = 1 , 2 , … , m ), select the best choice.

Step 6:

End.

Similar to the distance measure, we can apply the correlation coefficient to give a method to multi-attribute decision making. (<italic>See</italic> Bryson and Mobolurin, <xref ref-type="bibr" rid="j_info1142_ref_003">1995</xref>).

Let us consider the decision-making problem of evaluating university faculty for tenure and promotion. There are three faculty candidates (alternatives) A = { A 1 , A 2 , A 3 } to be evaluated using the linguistic term set S = { s 1 : extremely poor, s 2 : very poor, s 3 : poor, s 4 : slightly poor, s 5 : fair, s 6 : slightly good, s 7 : good, s 8 : very good, s 9 : extremely good} by an expert with respect to three attributes: C = { c 1 : teaching, c 2 : research, c 3 : service}. The associated assessment values are shown as listed in Table 1.

LHFSs of alternatives.

c 1 c 2 c 3
A 1 { ( s 5 , 0.1 , 0.2 ) , ( s 6 , 0.4 ) , s 7 , 0.3 ) } { ( s 6 , 0.4 ) , ( s 7 , 0.2 , 0.3 ) } { ( s 6 , 0.2 , 0.4 ) , ( s 7 , 0.3 ) }
A 2 { ( s 5 , 0.2 , 0.4 ) , ( s 6 , 0.3 , 0.5 ) } { ( s 7 , 0.3 , 0.6 ) , ( s 8 , 0.2 ) } { ( s 6 , 0.3 , 0.5 , 0.8 ) }
A 3 { ( s 5 , 0.2 ) , ( s 6 , 0.3 , 0.5 ) } { ( s 5 , 0.3 , 0.5 ) , ( s 6 , 0.2 , 0.3 ) , ( s 7 , 0.1 ) } { ( s 7 , 0.3 , 0.5 ) , ( s 8 , 0.1 , 0.3 ) }

These three faculty candidates are from one research-based university, which gives more importance to c 2 than to c 1 and c 3 , but, on the other hand, the committee gives some advantages to the candidates that are both good in c 2 and in either c 1 or c 3 . Their importance is defined as follows: μ ( c 1 ) = μ ( c 3 ) = 0.4 , μ ( c 2 ) = 0.6 , μ ( c 1 , c 2 ) = μ ( c 2 , c 3 ) = 0.85 , μ ( c 1 , c 3 ) = 0.7. Furthermore, the ordered position values’ importance is defined as follows: v ( 1 ) = 0.5 , v ( 2 ) = 0.6 , v ( 3 ) = 0.5 , v ( 1 , 2 ) = v ( 2 , 3 ) = 0.9 , v ( 1 , 3 ) = 0.8 , where N = { 1 , 2 , 3 }.

One can easily check that μ and v are both a 2-additive measure. Using the equation (15), the Shapley values of the attributes are ϕ c 1 ( C , μ ) = 0.275 , ϕ c 2 ( C , μ ) = 0.45, ϕ c 3 ( C , μ ) = 0.275; and the Shapley values of the ordered positions are ϕ 1 ( N , v ) = ϕ 3 ( N , v ) = 0.3 , ϕ 2 ( N , v ) = 0.4. To derive the optimal candidate, the following procedure is followed:

Because all attributes are benefit, we derive L H = L H ′ . According to the predefined linguistic term set S, we have M + = { ( s 9 , 1 ) , ( s 9 , 1 ) , ( s 9 , 1 ) } and M − = { ( s 1 , 0 ) , ( s 1 , 0 ) , ( s 1 , 0 ) }.

Using the H2SDM, we obtain:

H 2 SDM ( LH 1 , M + ) = 6.8328 , H 2 SDM ( LH 2 , M + ) = 5.8059 , H 2 SDM ( LH 3 , M + ) = 6.4055 and H 2 SDM ( LH 1 , M − ) = 1.5286 , H 2 SDM ( LH 2 , M − ) = 2.0305 , H 2 SDM ( LH 3 , M − ) = 1.2767.

From H 2 SDM ( LH i , M + ) and H 2 SDM ( LH i , M − ), we derive D 1 = 0.1828, D 2 = 0.2591, D 3 = 0.1662. Thus, the best candidate is A 2 .

In this example, when the GH2SWDM is applied, we have: GH 2 SDM ( LH 1 , M + ) = 6.8325 , GH 2 SDM ( LH 2 , M + ) = 5.7763 , GH 2 SDM ( LH 3 , M + ) = 6.3478 and GH 2 SDM ( LH 1 , M − ) = 1.5068 , GH 2 SDM ( LH 2 , M − ) = 1.9913 , GH 2 SDM ( LH 3 , M − ) = 1.2499 by which we get D 1 = 0.1807, D 2 = 0.2564, D 3 = 0.1645. It also shows that the candidate A 2 is the best choice.

Furthermore, when the H2SCC is used to select the best candidate, we obtain: H 2 SCC ( LH 1 , M + ) = 0.2417 , H 2 SCC ( LH 2 , M + ) = 0.3691 , H 2 SCC ( LH 3 , M + ) = 0.2615. From H 2 SCC ( LH 2 , M + ) > H 2 SCC ( LH 3 , M + ) > H 2 SCC ( LH 1 , M + )H2\mathit{SCC}({\mathit{LH}_{3}},{M^{+}})>H2\mathit{SCC}({\mathit{LH}_{1}},{M^{+}})$]]>, we know that the candidate A 2 is the best choice. When we adopt the GH2SCC to calculate the correlation coefficient, we derive: GH 2 SCC ( LH 1 , M + ) = 0.2407 , GH 2 SCC ( LH 2 , M + ) = 0.3489 , GH 2 SCC ( LH 3 , M + ) = 0.2538 which still shows that the candidate A 2 is the best choice.

Because the correlations between the attribute LHFSs of alternatives and the negative-ideal LHFS are zero, we only calculate the correlations between the attribute LHFSs of alternatives and the positive-ideal LHFS. Then, we give the ranking order of alternatives according to them.

 Comparison Analysis

Because most of previous studies applied multi-attribute decision making problem to linguistic hesitant fuzzy environment, several methods are applied to Example 2 to verify the effectiveness of the proposed approach.

Considering there are interactive characteristics between the attributes and their orders, Meng et al. (2014) proposed the GLHFHSWA and GLHFHSGM operators, by which the related results and ranking orders are obtained as shown in Tables 2 and 3.

Ranking results based on the GLHFHSWA operator to different values of γ.

E ( LH 1 ) E ( LH 2 ) E ( LH 3 ) Ranking orders
γ = 0.1 s 1.91 s 2.68 s 1.78 A 2 > A 1 > A 3 {A_{1}}>{A_{3}}$]]>
γ = 0.2 s 1.91 s 2.69 s 1.78 A 2 > A 1 > A 3 {A_{1}}>{A_{3}}$]]>
γ = 0.5 s 1.93 s 2.73 s 1.79 A 2 > A 1 > A 3 {A_{1}}>{A_{3}}$]]>
γ = 1.0 s 1.97 s 2.79 s 1.82 A 2 > A 1 > A 3 {A_{1}}>{A_{3}}$]]>
γ = 2.0 s 2.04 s 2.92 s 1.87 A 2 > A 1 > A 3 {A_{1}}>{A_{3}}$]]>
γ = 5.0 s 2.14 s 3.14 s 1.90 A 2 > A 1 > A 3 {A_{1}}>{A_{3}}$]]>
γ = 10.0 s 2.13 s 3.14 s 1.83 A 2 > A 1 > A 3 {A_{1}}>{A_{3}}$]]>
γ = 30.0 s 1.73 s 2.87 s 1.43 A 2 > A 1 > A 3 {A_{1}}>{A_{3}}$]]>

Ranking results based on the GLHFHSGM operator to different values of γ.

E ( LH 1 ) E ( LH 2 ) E ( LH 3 ) Ranking orders
γ = 0.1 s 1.89 s 2.45 s 1.62 A 2 > A 1 > A 3 {A_{1}}>{A_{3}}$]]>
γ = 0.2 s 1.89 s 2.44 s 1.61 A 2 > A 1 > A 3 {A_{1}}>{A_{3}}$]]>
γ = 0.5 s 1.88 s 2.41 s 1.61 A 2 > A 1 > A 3 {A_{1}}>{A_{3}}$]]>
γ = 1.0 s 1.88 s 2.37 s 1.59 A 2 > A 1 > A 3 {A_{1}}>{A_{3}}$]]>
γ = 2.0 s 1.86 s 2.29 s 1.56 A 2 > A 1 > A 3 {A_{1}}>{A_{3}}$]]>
γ = 5.0 s 1.82 s 2.11 s 1.48 A 2 > A 1 > A 3 {A_{1}}>{A_{3}}$]]>
γ = 10.0 s 1.74 s 1.94 s 1.38 A 2 > A 1 > A 3 {A_{1}}>{A_{3}}$]]>
γ = 30.0 s 1.58 s 1.74 s 1.20 A 2 > A 1 > A 3 {A_{1}}>{A_{3}}$]]>

Tables 2 and 3 show that the same ranking order and the same best choice are derived in this example, which is the same as that derived by using the H2SDM and GH2SWDM.

Most of previous studies assume that no interactive characteristics exist between the attributes and between the orders. To further show the advantage of the proposed approach under a linguistic hesitant fuzzy environment, we compare it with the methods proposed by Zhu et al. (2016) and Zhou et al. (2015) that are based on independent assumption. For the Zhu et al.’s method, the weights on the attributes are ω 1 = 0.275, ω 2 = 0.45 and ω 3 = 0.275, and the nine clouds are A 1 ( 0 , 3.01 , 0.107 ) , A 2 ( 1.93 , 2.75 , 0.193 ), A 3 ( 3.31 , 2.27 , 0.353 ), A 4 ( 4.30 , 1.932 , 0.466 ), A 5 ( 5 , 1.667 , 0.554 ), A 6 ( 5.70 , 1.932 , 0.466 ), A 7 ( 6.69 , 2.27 , 0.353 ), A 8 ( 8.07 , 2.75 , 0.193 ), A 9 ( 10 , 3.01 , 0.107 ) provided by Wang et al. (2014). The related results and ranking orders are shown in Table 4.

Ranking results based on the LHFPWA operator and the LHFPWG operator.

E ( LH 1 ) E ( LH 2 ) E ( LH 3 ) Ranking orders
LHFPWA s 1.97 s 2.72 s 1.78 A 2 > A 1 > A 3 {A_{1}}>{A_{3}}$]]>
LHFPWG s 1.88 s 2.32 s 1.56 A 2 > A 1 > A 3 {A_{1}}>{A_{3}}$]]>

Ranking results based on the ER method.

E ( LH 1 ) E ( LH 2 ) E ( LH 3 ) Ranking orders
Using LSF(1) s 0.46 s 0.44 s 0.42 A 1 > A 2 > A 3 {A_{2}}>{A_{3}}$]]>
Using LSF(2) s 0.43 s 0.41 s 0.40 A 1 > A 2 > A 3 {A_{2}}>{A_{3}}$]]>
Using LSF(3) s 0.48 s 0.45 s 0.42 A 1 > A 2 > A 3 {A_{2}}>{A_{3}}$]]>

With respect to the Zhou et al.’s extended evidential reasoning (ER) method, the related results and ranking orders are shown in Table 5.

Using three kinds of linguistic scale functions (LSFs) defined by Zhou et al. (2015), the ranking result is A 1 > A 2 > A 3 {A_{2}}>{A_{3}}$]]>, and the best alternative is A 1 .

From the above ranking results and ranking orders, one can find the difference between methods based on the independent and interactive analysis. This example shows that the new method is effective, and it is simpler than some previous methods.

 Conclusions

As we know, distance measure and correlation coefficient are two important tools to decision making. Considering the application of linguistic hesitant fuzzy sets, this paper defines a distance measure, and then introduces a correlation coefficient. After that, we develop two types of the hybrid weighted distance measures and the hybrid weighted correlation coefficients for linguistic hesitant fuzzy sets, by which we can derive the comprehensive evaluated values of the objects. Furthermore, we study their applications to pattern recognition and to multi-attribute decision making.

Comparing with the previous researches about decision making with LHFSs, there are several contributions of our method:

It is simpler than the previous method (Meng et al., 2014);

It addresses the issues in the previous distance measure (Zhou et al., 2015);

It considers the interaction between elements in a set and the complexity of determining a fuzzy measure;

It extends the application of LHFSs.

 However, we only present one distance measure and one correlation coefficient, and it will be interesting to study other distance measures and correlation coefficients for linguistic hesitant fuzzy sets. Furthermore, we shall discuss their application in other fields, such as expert systems, digital image processing, and clustering analysis. Moreover, we can extend the developed theoretical results to other types of fuzzy sets, such as hesitant interval neutrosophic linguistic sets (Ye, 2013).

 Acknowledgements

This work was supported by the State Key Program of National Natural Science of China (No. 71431006), the Projects of Major International Cooperation NSFC (No. 71210003), the National Natural Science Foundation of China (Nos. 71571192 and 71271080), the Ministry of Education Humanities Social Science Foundation of China (13YJA630020), the Research Foundation of Education Bureau of Hunan Province, China (16C0515), and the Innovation-Driven Planning Foundation of Central South University (Nos. 2015CX010, 2016CXS027).

 References Atanassov, K. (1986). Intuitionistic fuzzy sets. Fuzzy Sets and Systems, 20, 87–96. Atanassov, K., Gargov, G. (1989). Interval valued intuitionistic fuzzy sets. Fuzzy Sets and Systems, 31, 343–349. Bryson, N., Mobolurin, A. (1995). An action learning evaluation procedure for multiple criteria decision making problems. European Journal of Operational Research, 96, 379–386. Cabrerizo, F.J., Morente-Molinera, J.A., Pérez, I.J., López-Gijón, J., Herrera-Viedma, E. (2015). A decision support system to develop a quality management in academic digital libraries. Information Sciences, 323, 48–58. Cai, M., Gong, Z.W., Wu, D.Q., Wu, M.J. (2014a). A pattern recognition method based on linguistic ordered weighted distance measure. Journal of Intelligent and Fuzzy Systems, 27, 1897–1903. Cai, M., Gong, Z.W., Cao, J., Wu, M.J. (2014b). A novel distance measure of multi-granularity linguistic variables and its application to MADM. International Journal of Fuzzy Systems, 16, 378–388. Cai, M., Gong, Z.W., Cao, J. (2015). The consistency measures of multi-granularity linguistic group decision making. Journal of Intelligent and Fuzzy Systems, 29, 609–618. Chen, L.H., Weng, M.C. (2006). An evaluation approach to engineering design in QFD processes using fuzzy goal programming models. European Journal of Operational Research, 172, 230–248. Chen, N., Xu, Z.S., Xia, M.M. (2013a). Interval-valued hesitant preference relations and their applications to group decision making. Knowledge-Based Systems, 37, 528–540. Chen, N., Xu, Z.S., Xia, M.M. (2013b). Correlation coefficients of hesitant fuzzy sets and their applications to clustering analysis. Applied Mathematical Modelling, 37, 2197–2211. Dong, Y.C., Xu, Y.F., Yu, S. (2009). Computing the numerical scale of the linguistic term set for the 2-tuple fuzzy linguistic representation model. IEEE Transactions on Fuzzy Systems, 17, 1366–1378. Dong, Y.C., Li, C.C., Herrera, F. (2016). Connecting the linguistic hierarchy and the numerical scale for the 2-tuple linguistic model and its use to deal with hesitant unbalanced linguistic information. Information Sciences, 367, 259–278. Fu, G.T. (2008). A fuzzy optimization method for multicriteria decision making: An application to reservoir flood control operation. Expert Systems with Applications, 34, 145–149. Gong, Z.W., Xu, X.X., Yang, Y.J., Zhou, Y., Zhang, H.H. (2016). The spherical distance for intuitionistic fuzzy sets and its application in decision analysis. Technological and Economic Development of Economy, 22, 393–415. Gou, X.J., Xu, Z.S. (2016). Novel basic operational laws for linguistic terms, hesitant fuzzy linguistic term sets and probabilistic linguistic term sets. Information Sciences, 372, 407–427. Grabisch, M. (1997). k-order additive discrete fuzzy measures and their representation. Fuzzy Sets and Systems, 92, 167–189. Herrera, F., Martínez, L. (2000). A 2-tuple fuzzy linguistic representation model for computing with words. IEEE Transactions on Fuzzy Systems, 8, 746–752. Herrera, F., Herrera-Viedma, E., Martínez, L. (2000). A fusion approach for managing multi-granularity linguistic term sets in decision making. Fuzzy Sets and Systems, 114, 43–58. Hung, W.L., Yang, M.S. (2004). Similarity measures of intuitionistic fuzzy sets based on Hausdorff distance. Pattern Recognition Letters, 25, 1603–1611. James, G., Dolan, M.D. (2010). Multi-criteria clinical decision support: a primer on the use of multiple criteria decision making methods to promote evidence-based, patient-centered healthcare. The Patient: Patient-Centered Outcomes Research, 3, 229–248. Ju, Y.B., Liu, X.Y., Ju, D.W. (2016). Some new intuitionistic linguistic aggregation operators based on Maclaurin symmetric mean and their applications to multiple attribute group decision making. Soft Computing, 20, 4521–4548. Kahraman, C., Ruan, D., Doǧan, I. (2003). Fuzzy group decision-making for facility location selection. Information Sciences, 157, 135–153. Lennon, E., Farr, J., Besser, R. (2013). Evaluation of multi-attribute decision making systems applied during the concept design of new microplasma devices. Expert Systems with Applications, 40, 6321–6329. Li, C.C., Dong, Y.C., Herrera, F., Herrera-Viedma, E., Martínez, L. (2017). Personalized individual semantics in Computing with Words for supporting linguistic group decision making. An application on consensus reaching. Information Fusion, 33, 29–40. Liang, Z., Shi, P. (2003). Similarity measures on intuitionistic fuzzy sets. Pattern Recognition Letters, 24, 2687–2693. Liu, P.D. (2013). Some geometric aggregation operators based on interval intuitionistic uncertain linguistic variables and their application to group decision making. Applied Mathematical Modelling, 37, 2430–2444. Liu, P.D., Jin, F. (2012). Methods for aggregating intuitionistic uncertain linguistic variables and their application to group decision making. Information Sciences, 205, 58–71. Martínez, L., Herrera, F. (2012). An overview on the 2-tuple linguistic model for computing with words in decision making: extensions, applications and challenges. Information Sciences, 207, 1–18. Martínez-Cruz, C., Porcel, C., Bernabé-Moreno, J., Herrera-Viedma, E. (2015). A model to represent users trust in recommender systems using ontologies and fuzzy linguistic modeling. Information Sciences, 311, 102–118. Massanet, S., Riera, J.V., Torrens, J., Herrera-Viedma, E. (2014). A new linguistic computational model based on discrete fuzzy numbers for computing with words. Information Sciences, 258, 277–290. Meng, F.Y., An, Q.X. (2017). An approach for group decision making method with hesitant fuzzy preference relations. Knowledge-Based Systems. https://doi.org/10.1016/j.knosys.2017.03.010. Meng, F.Y., Chen, X.H. (2015). Correlation coefficients of hesitant fuzzy sets and their application based on fuzzy measures. Cognitive Computation, 7, 445–463. Meng, F.Y., Chen, X.H. (2016a). Entropy and similarity measure of Atanassov’s intuitionistic fuzzy sets and their application to pattern recognition based on fuzzy measures. Pattern Analysis and Applications, 19, 11–20. Meng, F.Y., Chen, X.H. (2016b). The symmetrical interval intuitionistic uncertain linguistic operators and their application to decision making. Computers and Industrial Engineering, 98, 531–542. Meng, F.Y., Chen, X.H. (2016c). Entropy and similarity measure for Atannasov’s interval-valued intuitionistic fuzzy sets and their applications. Fuzzy Optimization and Decision Making, 15, 75–101. Meng, F.Y., Chen, X.H. (2017a). A new method for triangular fuzzy compare wise judgement matrix process based on consistency analysis. International Journal of Fuzzy Systems, 19, 27–46. Meng, F.Y., Chen, X.H. (2017b). Correlation coefficient of interval-valued intuitionistic uncertain linguistic sets and its application. Cybernetics and Systems, 48, 114–135. Meng, F.Y., Tang, J. (2013). Interval-valued intuitionistic fuzzy multi-criteria group decision making based on cross entropy and Choquet integral. International Journal of Intelligent Systems, 28, 1172–1195. Meng, F.Y., Chen, X.H., Zhang, Q. (2014). Multi-attribute decision analysis under a linguistic hesitant fuzzy environment. Information Sciences, 267, 287–305. Meng, F.Y., An, Q.X., Chen, X.H. (2016a). A consistency and consensus-based method to group decision making with interval linguistic preference relations. Journal of the Operational Research Society, 67, 1419–1437. Meng, F.Y., Wang, C., Chen, X.H., Zhang, Q. (2016b). Correlation coefficients of interval-valued hesitant fuzzy sets and their application based on Shapley function. International Journal of Intelligent Systems, 31, 17–43. Meng, F.Y., Zhang, Q., Chen, X.H. (2016c). Fuzzy multichoice games with fuzzy characteristic functions. Group Decision and Negotiation. https://doi.org/10.1007/s10726-016-9493-7. Meng, F.Y., An, Q.X., Tan, C.Q., Chen, X.H. (2016d). An approach for group decision making with interval fuzzy preference relations based on additive consistency and consensus analysis. IEEE Transactions on Systems, Man, and Cybernetics Systems. https://doi.org/10.1109/TSMC.2016.2606647. Meng, F.Y., Tang, J., An, Q.X., Chen, X.H. (2016e). Decision making with intuitionistic linguistic preference relations. International Transactions in Operational Research. https://doi.org/10.1111/itor.12383. Meng, F.Y., Zhou, D., Chen, X.H. (2016f). An approach to hesitant fuzzy group decision making with multi-granularity linguistic information. Informatica, 27, 767–798. Meng, F.Y., Lin, J., Tan, C.Q., Zhang, Q. (2017a). A new multiplicative consistency based method for decision making with triangular fuzzy reciprocal preference relations. Fuzzy Sets and Systems, 315, 1–25. Meng, F.Y., Tan, C.Q., Chen, X.H. (2017b). Multiplicative consistency analysis for interval reciprocal preference relations: a comparative study. Omega, 68, 17–38. Merigó, J.M., Casanovas, M. (2011). Induced aggregation operators in the Euclidean distance and its application in financial decision making. Expert Systems with Applications, 38, 7603–7608. Morente-Molinera, J.A., Pérez, I.J., Ureña, M.R., Herrera-Viedma, E. (2015). On multi-granular fuzzy linguistic modeling in group decision making problems: a systematic review and future trends. Knowledge-Based Systems, 74, 49–60. Ölçer, A., Tuzcu, C., Turan, O. (2006). An integrated multi-objective optimisation and fuzzy multi-attributive group decision-making technique for subdivision arrangement of Ro-Ro vessels. Applied Soft Computing, 6, 221–243. Park, D.G., Kwun, Y.C., Park, J.H., Park, I.Y. (2009). Correlation coefficient of interval-valued intuitionistic fuzzy sets and its application to multiple attribute group decision making problems. Mathematical and Computer Modelling, 50, 1279–1293. Pedrycz, W. (2013). Granular Computing: Analysis and Design of Intelligent Systems. CRC Press, Boca Raton. Peng, D.H., Gao, C.Y., Gao, Z.F. (2013). Generalized hesitant fuzzy synergetic weighted distance measures and their application to multiple criteria decision-making. Applied Mathematical Modelling, 37, 5837–5850. Rodríguez, R.M., Martinez, L., Herrera, F. (2012). Hesitant fuzzy linguistic term sets for decision making. IEEE Transactions on Fuzzy Systems, 20, 109–119. Shapley, L.S. (1953). A Value for n-Person Game. Princeton University Press, Princeton. Sugeno, M. (1974). Theory of Fuzzy Integral and Its Application. Doctoral dissertation, Tokyo Institute of Technology. Tan, C.Q. (2011). A multi-criteria interval-valued intuitionistic fuzzy group decision making with Choquet integral-based TOPSIS. Expert Systems with Applications, 38, 3023–3033. Tan, C.Q., Chen, X.H. (2011). Induced intuitionistic fuzzy Choquet integral operator for multi-criteria decision making. International Journal of Intelligent Systems, 26, 659–686. Tejeda-Lorente, Porcel C, A., Peis, E., Sanz, R., Herrera-Viedma, E. (2014). A quality based recommender system to disseminate information in a university digital library. Information Sciences, 261, 52–69. Tong, X., Yu, L. (2016). MADM based on distance and correlation coefficient measures with decision-maker preferences under a hesitant fuzzy environment. Soft Computing, 20, 4449–4461. Torra, V. (2010). Hesitant fuzzy sets. International Journal of Intelligent Systems, 25, 529–539. Tran, L., Duckstein, L. (2002). Comparison of fuzzy numbers using a fuzzy distance measure. Fuzzy Sets and Systems, 130, 331–341. Vaidogas, E.R., Sakenaite, J. (2011). Multi-attribute decision-making in economics of fire protection. Engineering Economics, 22, 262–270. Van der Weken, D., Nachtegael, M., Kerre, E.E. (2004). Using similarity measures and homogeneity for the comparison of images. Image and Vision Computing, 22, 695–702. Wang, J.Q., Li, J.J. (2009). The multi-criteria group decision making method based on multi-granularity intuitionistic two semantics. Science and Technology Information, 33, 8–9. Wang, J.Q., Lu, P., Zhang, H.Y., Chen, X.H. (2014). Method of multi-criteria group decision-making based on cloud aggregation operators with linguistic information. Information Sciences, 274, 177–191. Wei, G.W. (2011). Some generalized aggregating operators with linguistic information and their application to multiple attribute group decision making. Computers and Industrial Engineering, 61, 32–38. Wei, G.W., Wang, H.J., Lin, R. (2011). Application of correlation coefficient to interval-valued intuitionistic fuzzy multiple attribute decision-making with incomplete weight information. Knowledge and Information Systems, 26, 337–349. Wu, Z.B., Xu, J.P. (2016). Managing consistency and consensus in group decision making with hesitant fuzzy linguistic preference relations. Omega, 65, 28–40. Xu, Z.S. (2004a). A method based on linguistic aggregation operators for group decision making with linguistic preference relations. Information Sciences, 166, 19–30. Xu, Z.S. (2004b). Uncertain linguistic aggregation operators based approach to multiple attribute group decision making under uncertain linguistic environment. Information Sciences, 168, 171–184. Xu, Z.S. (2007). A method for multiple attribute decision making with incomplete weight information in linguistic setting. Knowledge-Based Systems, 20, 719–725. Xu, Z.S. (2010a). Choquet integrals of weighted intuitionistic fuzzy information. Information Sciences, 180, 726–736. Xu, Z.S. (2010b). A method based on distance measure for interval-valued intuitionistic fuzzy group decision making. Information Sciences, 180, 181–190. Xu, Z.S., Chen, J. (2008). Ordered weighted distance measure. Journal of Systems Science and Systems Engineering, 17, 432–445. Xu, Z.S., Xia, M.M. (2011). Distance and similarity measures for hesitant fuzzy sets. Information Sciences, 181, 2128–2138. Xu, Z.S., Chen, J., Wu, J.J. (2008). Clustering algorithm for intuitionistic fuzzy sets. Information Sciences, 178, 3775–3790. Yager, R.R. (1988). On ordered weighted averaging aggregation operators in multi-criteria decision making. IEEE Transactions on Systems, Man and Cybernetics, 18, 183–190. Yager, R.R. (2003). Fuzzy Logic methods in recommender systems. Fuzzy Sets and Systems, 136, 133–149. Yager, R.R. (2004). Customization using fuzzy recommender systems. Journal of Donghua University, 21, 9–14. Yang, M.S., Lin, D.C. (2009). On similarity and inclusion measures between type-2 fuzzy sets with an application to clustering. Computers and Mathematics with Applications, 57, 896–907. Ye, J. (2013). Multiple attribute decision-making method under hesitant interval neutrosophic linguistic environment. Zenodo. https://doi.org/10.5281/zenodo.34858. Ye, J. (2015). Multiple-attribute decision-making method under a single-valued neutrosophic hesitant fuzzy environment. Journal of Intelligent Systems, 24, 23–36. Ye, J. (2016a). Aggregation operators of neutrosophic linguistic numbers for multiple attribute group decision making. SpringerPlus, 5, 1691. Ye, J. (2016b). Correlation coefficients of interval neutrosophic hesitant fuzzy sets and their multiple attribute decision making method. Informatica, 27, 179–202. Yue, Z.L. (2011). Deriving decision maker’s weights based on distance measure for interval-valued intuitionistic fuzzy group decision making. Expert Systems with Applications, 38, 11665–11670. Zadeh, L.A. (1965). Fuzzy sets. Information Control, 8, 338–353. Zadeh, L.A. (1973). Outline of a new approach to the analysis of complex systems and decision processes interval-valued fuzzy sets. IEEE Transactions on Systems, Man and Cybernetics, SMC-3, 28–44. Zadeh, L.A. (1975). The concept of a linguistic variable and its application to approximate reasoning part I. Information Sciences, 8, 199–249. Zeng, S.Z., Su, W.H. (2011). Intuitionistic fuzzy ordered weighted distance operator. Knowledge-Based Systems, 24, 1224–1232. Zeng, W.Y., Li, D.Q., Yin, Q. (2016). Distance and similarity measures between hesitant fuzzy sets and their application in pattern recognition. Pattern Recognition Letters, 84, 267–271. Zhang, Q.S., Jiang, S.Y. (2008). A note on information entropy measures for vague sets and its applications. Information Sciences, 178, 4184–4191. Zhang, Y., Zhou, G., Jin, J., Wang, X., Cichocki, A. (2013). L1-regularized multiway canonical correlation analysis for SSVEP-based BCI. IEEE Transactions on Neural Systems and Rehabilitation, 21, 887–896. Zhang, Y., Zhou, G., Jin, J., Wang, X., Cichocki, A. (2014). Frequency recognition in SSVEP-based BCI using multiset canonical correlation analysis. International Journal of Neural Systems, 24, 1450013. Zhou, H., Wang, J.Q., Zhang, H.Y., Chen, X.H. (2015). Linguistic hesitant fuzzy multi-criteria decision-making method based on evidential reasoning. International Journal of Systems Science, 47, 314–327. Zhu, C.X., Zhu, L., Zhang, X.Z. (2016). Linguistic hesitant fuzzy power aggregation operators and their applications in multiple attribute decision-making. Information Sciences, 367, 809–826.