Scholarly article on topic 'Measurement of Job Satisfaction Using Fuzzy Sets'

Measurement of Job Satisfaction Using Fuzzy Sets Academic research paper on "Economics and business"

CC BY-NC-ND
0
0
Share paper
Academic journal
Procedia Computer Science
OECD Field of science
Keywords
{"Job satisfaction" / "fuzzy conjoint model" / "similarity degree."}

Abstract of research paper on Economics and business, author of scientific article — Rahib H. Abiyev, Tulen Saner, Serife Eyupoglu, Gunay Sadikoglu

Abstract The measurement of job satisfaction is carried out using the set of variables of some criterions defined by the experts. This measurement is basically based on linguistic evaluations of the variables. In the paper these linguistic values are represented by fuzzy sets as preference levels. A fuzzy set theory was combined with conjoint analysis and used for measurement of the job satisfaction of hotel employees. The algorithm for solving of fuzzy job satisfaction has been designed. The attributes for evaluation of job satisfaction are determined and employees’ opinions were recorded in the form of preference degrees. The statistical data describing satisfaction levels of hotel employees are collected. The membership degrees for each attribute are determined and using a similarity measure the closest of opinions of the employees and experts’ are determined. These closest degrees are ranked, and the evaluation of job satisfaction has been performed.

Academic research paper on topic "Measurement of Job Satisfaction Using Fuzzy Sets"

CrossMark

Available online at www.sciencedirect.com

ScienceDirect

Procedía Computer Science 102 (2016) 294 - 301

12th International Conference on Application of Fuzzy Systems and Soft Computing, ICAFS

2016, 29-30 August 2016, Vienna, Austria

Measurement of job satisfaction using fuzzy sets

Rahib H.Abiyeva*, Tulen Sanerb, Serife Eyupogluc, Gunay Sadikoglub

a Department of Computer Engineering, Near East University, P.O.BOX:99138, Nicosia, North Cyprus, Mersin 10 Turkey bSchool of Tourism and Hotel Management, Near East University, P.O.BOX:99138, Nicosia, North Cyprus, Mersin 10, Turkey cDepartment of Business Administration, Near East University, P.O.BOX:99138, Nicosia, North Cyprus, Mersin 10, Turkey

Abstract

The measurement of job satisfaction is carried out using the set of variables of some criterions defined by the experts. This measurement is basically based on linguistic evaluations of the variables. In the paper these linguistic values are represented by fuzzy sets as preference levels. A fuzzy set theory was combined with conjoint analysis and used for measurement of the job satisfaction of hotel employees. The algorithm for solving of fuzzy job satisfaction has been designed. The attributes for evaluation of job satisfaction are determined and employees' opinions were recorded in the form of preference degrees. The statistical data describing satisfaction levels of hotel employees are collected. The membership degrees for each attribute are determined and using a similarity measure the closest of opinions of the employees and experts' are determined. These closest degrees are ranked, and the evaluation of job satisfaction has been performed.

© 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.Org/licenses/by-nc-nd/4.0/).

Peer-review under responsibility of the Organizing Committee of ICAFS 2016 Keywords: Job satisfaction; fuzzy conjoint model; similarity degree.

1. Introduction

Recently a set of research works have been done on job satisfaction. Some of these studies use percentage and mean values for the analysis. Likert scale is used to study and evaluate job satisfaction of library staff at the University of North California3. The paper4 uses Likert scale for evaluation of job satisfaction in South African University. The paper5 uses five and seven points Likert scale to evaluate job satisfaction by analysing the measures of statistical mean, standard deviation and correlation. In6 gives analyse of job satisfaction of college teachers. The

* Rahib H.Abiyev. Tel.: +090-392-2236464; fax: +090-392-2236622. E-mail address.Tahib.abiyev@neu.edu.tr

1877-0509 © 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.Org/licenses/by-nc-nd/4.0/).

Peer-review under responsibility of the Organizing Committee of ICAFS 2016 doi:10.1016/j.procs.2016.09.404

use of artificial intelligence techniques in assessment and evaluation of performances of teachers has been considered in research papers7'8 .

The analysis of job satisfaction of hotel employees is important in service business. Hotel enterprise is as an important part of hospitality. The level of job satisfaction of hotel employees has influence to their behaviour, and then has influence to the customer's satisfaction directly. Consequently the job satisfaction of employees affects the efficiency of the hotels. The satisfaction of employees increases their retention, productivity, a higher level of service quality9. There are few studies on job satisfaction of the hotel employees. The relationship between demographic characteristics of hotel employees and job satisfaction has been examined in10'11'12. These researches are basically based on statistical analysis.

Fuzzy set theory provides an excellent framework for describing imprecise meaning of preferences and their subjective nature. Recently a set of research works have been done for evaluation of job satisfaction using fuzzy set theory. The paper13 used fuzzy set theory for evaluation of job satisfaction of academic staff. Fuzzy sets theory is also used for evaluation of students' perceptions on computer algebra14, teaching quality15, teachers' beliefs16, credit card services17. In these researches, different approaches have been used for solution of job satisfaction problem.

The paper is organized as follows. Sec.2 gives descriptions of the fuzzy sets and fuzzy conjoint analysis. The conjoint model used for evaluation of job satisfaction of hotel employees is presented. Sec. 3 describes the experimental results. The flowchart of the algorithm used for evaluation of job satisfaction of hotel employees using conjoint analysis is presented. Sec. 4 includes comparative results of existing methods. Finally, in Sec.5 the conclusions are presented.

2. Fuzzy conjoint model for measurement of job satisfaction

Conjoint analysis was used to study preferred levels of individuals and relative importance of the multiple attributes of market goods18. Individuals can evaluate the multi- attributes using responses that are approximately interval in a measurement level. The requirements of one group may have a conflict with the requirements of other groups and evaluation these factors among diverse customer needs is critical. In our case, a multivariate technique is used to demonstrate how respondents develop preferences. Conjoint analysis allows to estimate the relative importance of selected attributes. Respondents express their preferences by providing the importance of the attribute.

The preferences of respondents are basically vague, uncertain and they have a subjective nature based on feeling individuals. Fuzzy set theory provides an excellent framework for describing preferences, their subjective nature. Due to subjective uncertainty, imprecise meaning of preferences, fuzzy set theory was combined with conjoint analysis. The variables used in job satisfaction are linguistic variables and their values are estimated by preferences of individuals. For example: satisfied, very satisfied, unsatisfied etc.

Fuzzy conjoint analysis proposed by Turksen and Willson (1994) is used for the analysis of consumer preferences in marketing18. Fuzzy sets are used to represent the values of the attributes evaluated by respondents. The membership degree of element yj for the linguistic label representing item A is defined as

satisfaction, (Xj, A) is the membership degree for respondent j for item A according linguistic label

xj=1,2,.. ^n, n is a number of linguistic term, A is an item/a question.

The membership degree represents the fuzzy set of response of the respondent. This fuzzy set is compared with fuzzy set defined by expert. The comparison is done using fuzzy similarity measure which is based on Euclidian distance of two fuzzy sets. The similarity measure is calculated as follows.

Mr (y j , A) = 2 -— " ^F (Xj , A)

Wi is the weight that represent level of

Sim(R (yf, A), F(xf, l)) -

E (^R.( yj, A)-^f (Xj, l ))2

i=1,..,M, j=1,-,N

where R (yj, A) is the fuzzy sets determined using the responses of respondents, F(Xj, l) is the standard fuzzy sets

determined for linguistic label l. M is the number of attributes, N is the number of linguistic terms. The similarity is computed for product m for each of the n possible linguistic terms. The similarity rate ranges from 0 to 1. Here N is the number of members in the linguistic variable vector.

3. Experimental results

In order to ascertain the job satisfaction of employees, we requested to specify their level of satisfaction or dissatisfaction using Likert's 5 point Scale. Each of the questions was evaluated by linguistic terms. In the paper, five linguistic terms with five satisfaction level were created. The scale was represented as follows: Very Satisfied (VS), Satisfied (S), Neutral (N), Dissatisfied (DS) and Very Dissatisfied (VDS). Fuzzy set theory is applied to represent these linguistic variables. The triangular membership functions are used to represent the fuzzy sets. Fig. 1 illustrates an example of the fuzzy sets representing the level of satisfaction. The membership functions of the fuzzy terms are represented by following formula.

a3 - X

a - a2

if a - x ^ a2 if a2 < x < a

otherwise

where a2 is centre, a1 is left and a3 is right sides of triangle.

As mentioned above 20 attributes are applied to the questionnaire. The respondents were selected one of the answers of these variables in each choice. The level of satisfaction of the employees for each question in terms of 5 point Likert's scales is given in Table 1. The numbers of the employees and their percentages are given in the table. The results of answers for item 1 and item 2 are reported below.

- Item 1. Activity: 13% of the employees were very dissatisfied with the activities in the hotels, 13%- dissatisfied, while 42% of the employees were satisfied, 15% - very satisfied and 17% of the employees preferred to be neutral.

- Item 2. Independence: 11% of the employees were very dissatisfied with the way that they are allowed to work independently, 14% of the employees were dissatisfied, while 41% of the employees expressed that they were satisfied, 15%- very satisfied. 19% of the employees were neutral.

Similarly, the results of answers for other attributes could be analysed.

Membership

Fig. 1. The fuzzy sets representing the level of satisfaction.

Rahib H. Abiyev et al. /Procedia Computer Science 102 (2016) 294 — 301 Table 1. Job Satisfaction Level Analysis

Scale VDS DS N S VS

Ques. Fr. % Fr. % Fr. % Fr. % Fr. %

Item 1 31 13 31 13 44 17 106 42 36 15

Item 2 28 11 35 14 47 19 100 41 38 15

Item 20 23 9 20 8 28 11 91 37 86 35

The input data given in Table 1 were provided to the input of job satisfaction system. Preference levels are determined using linguistic terms 'very satisfied', 'satisfied', 'neutral', 'dissatisfied', 'very dissatisfied'. According to the number of linguistic terms, the universe of fuzzy set is defined as (k= 1, 2, 3, 4, 5).

The five membership functions are defined using following expressions. "Very Satisfied" is defined as

F =i V 0.6/ 0 2/ 0/ 0/) , "Satisfied"- F =/0.6/ 1/ 0.6/ 0.2/ 0/) , "Neutral"- F =/0.2/ 0.6/ / 0.6/ 0.2/) , F 1/1' /2' /3'/4'/5J ' f2 \ 7172' /3, /4v5J F j /1 , /2,/з, AT /5J '

"Dissatisfied"- F =/0/ 0.2/ 0.6/ / 0.6/), "Very Dissatisfied"- F =/0/ 0/ 0.2/ 0.6/ /) F 1/1 , /2, /3 V4, /5] J F 1/1 V21 /3, /47M

In the second step using attributes and preference levels the questionnaire for job satisfaction problem is constructed and distributed among hotel employees. Hotel employees' opinions for each attribute in the questionnaire are collected. 248 questionnaires are collected from the employees of different hotels of North Cyprus. The collected questionnaires are analyzed using employees' opinions regarding to a selected linguistic variable. Table 1 demonstrates the results of the analysis. For example, in the table for the first item, 31 employees had chosen very dissatisfied (13% of employees), 31 (13%) dissatisfied, 44(14%) neutral, 106 (42%) satisfied, 36 (15%) very satisfied. Analysis has been performed for each attribute of questionnaires. After analysis of questionnaire results, in third step, the satisfaction degree levels for each attribute is calculated using fuzzy CA model of Turksen and Willson's formula (1). The operations have been performed by computing weight and correspondingly the membership degree (R) for respondent j for the item A according linguistic label. Table 2 depicts the values of membership functions. After calculating membership degrees in the fourth step the similarity degree values between employees' opinions and experts' opinions are calculated. This operation has been done by computing similarity degree between fuzzy sets R and F using formula (2). In last fifth step, the maximum amount of similarity degrees for each state is determined. Similarity degree demonstrates the maximum closeness of experts and customers' opinions to each other. Table 3 demonstrates the similarity degrees between fuzzy sets F and R.

Table 2. The Values of membership degree of fuzzy sets R.

F1 F2 F3 F4 F5

R1 0.1250 0.1250 0.1774 0.4274 0.1452

R2 0.1129 0.1411 0.1895 0.4032 0.1532

R3 0.1129 0.1290 0.1573 0.4355 0.1653

R4 0.0887 0.1089 0.1169 0.3750 0.3105

R5 0.0847 0.0766 0.1331 0.4274 0.2782

R6 0.0403 0.1411 0.1492 0.3790 0.2903

R7 0.0887 0.1008 0.1815 0.3105 0.3185

R8 0.0645 0.0605 0.1613 0.4677 0.2460

R9 0.0605 0.0887 0.1694 0.4153 0.2661

R10 0.0927 0.0968 0.1492 0.4153 0.2460

R11 0.0847 0.0927 0.1532 0.4032 0.2661

R12 0.1331 0.1411 0.2742 0.3024 0.1492

R13 0.1734 0.1694 0.2460 0.2863 0.1250

R14 0.1371 0.1210 0.1815 0.3911 0.1694

R15 0.1169 0.1290 0.1653 0.4355 0.1532

R16 0.0887 0.0847 0.1411 0.4637 0.2218

R17 0.0806 0.1089 0.1855 0.4395 0.1855

R18 0.0887 0.1089 0.1573 0.3952 0.2500

R19 0.1331 0.0927 0.1653 0.3266 0.2823

R20 0.0927 0.0806 0.1129 0.3669 0.3468

After calculating similarity degrees, the maximum similarity degrees for each attribute among all states are

selected. Ranks are designated for the selected maximum values. Ranking is based maximum similarity degree among all states. The results of the selection of the maximum similarity degrees and results of ranking are shown in Table 4.

Table 3. The Values of similarity degree between

Table 4. Maximum similarity degree and ranking Maximum Ranking Linguistic values

F, F2 F3 F4 F5 similarity degree for each state

R1 0 5385 6 satisfied

R1 0.4777 0.4729 0.5076 0 5385 0 5291 R2 0 5375 6 satisfied

R2 0.4789 0.4764 0.5110 0 5375 0 5297 R3 0 5407 5 satisfied

R3 0.4744 0.4697 0.5041 0 5407 0 5351 R4 0 5736 2 very satisfied

R4 0.4658 0.4566 0.4885 0 5403 0 5736 R5 0 5702 2 very satisfied

R5 0.4602 0.4524 0.4910 0 5496 0 5702 R6 0 5684 3 very satisfied

R6 0.4615 0.4607 0.4979 0 5463 0 5684 R7 0 5705 2 very satisfied

R7 0.4694 0.4617 0.4964 0 5363 0 5705 R8 0 5645 3 very satisfied

R8 0.4539 0.4500 0.4960 0 5585 0 5645 R9 0 5666 3 very satisfied

R9 0.4591 0.4559 0.4988 0 5527 0 5666 R10 0 5588 4 very satisfied

R10 0.4661 0.4595 0.4969 0 5463 0 5588 R11 0 5645 3 very satisfied

R11 0.4643 0.4580 0.4962 0 5470 0 5645 R12 0 5252 7 satisfied

R12 0.4902 0.4887 0.5228 0 5252 0 5195 R13 0 5190 8 neutral

R13 0.5040 0.4975 0.5190 0 5144 0 5098 R14 0 5348 6 satisfied

R14 0.4819 0.4744 0.5067 0 5348 0 5330 R15 0 5400 5 satisfied

R15 0.4756 0.4712 0.5058 0 5400 0 5317 R16 0 5550 4 very satisfied

R16 0.4613 0.4555 0.4957 0 5513 0 5550 R17 0 5494 5 satisfied

R17 0.4659 0.4647 0.5071 0 5494 0 5431 R18 0 5583 4 very satisfied

R18 0.4680 0.4624 0.4990 0 5444 0 5583 R19 0 5600 3 very satisfied

R19 R20 0.4776 0.4621 0.4648 0.4505 0.4951 0.4832 0 0 .5322 5401 0 0 5600 5849 R20 0 5849 1 very satisfied

As can be seen in Table 4, the results are obtained with 5% neutral, 35% satisfied, and 60% very satisfied. As can be seen in Table 4, the 20-th state- achievement with 9% very dissatisfied, 6% dissatisfied, 11% neutral, 37% satisfied, and 35% very satisfied has a best rank. The states 4 (with 8.8% very dissatisfied, 11% dissatisfied, 11.7% neutral, 37.5% satisfied and 31.5% very satisfied), 5 (with 8.5% very dissatisfied, 7.7% dissatisfied, 13.3% neutral, 42.7% satisfied and 27.8% very satisfied) and 7 (with 8.9% very dissatisfied, 10% dissatisfied, 18.1% neutral, 31% satisfied and 32% very satisfied) are ranked as the second. The worst states are 13 (with 17.3% very dissatisfied, 17% dissatisfied, 24.5% neutral, 28.6% satisfied and 12.5% very satisfied), 12 (with 13.3% very dissatisfied, 14% dissatisfied, 27.5% neutral, 30.2% satisfied and 15% very satisfied), 14 (with 13.7% very dissatisfied, 12.1% dissatisfied, 28.1% neutral, 39% satisfied and 17% very satisfied), 2 (with 11.3% very dissatisfied, 14.1% dissatisfied, 19% neutral, 40.3% satisfied and 15.3% very satisfied) and 1 (with 12.5% very dissatisfied, 12.5% dissatisfied, 17.7% neutral, 42.7.2% satisfied and 14.5% very satisfied). Since input parameters of model are fuzzy variables, output of proposed model (rankings) would be fuzzy

We apply a method of non-parametrical correlation to measure the statistical relation of ranks of alternatives21,22. Comparisons of the approaches are based on Spearman's rank correlation test. Spearman's rank correlation coefficient can be used to represent the degree of dependence between two variables. It is calculated from the ranks provided with every pair of the applied methods. Spearman's rank correlation coefficient rs can be calculated applying the following equation21:

The paper23 used Spearman's rank correlation to compare the ranks achieved by multiple criteria decision making methods. Using (4) and (5) Z statistics is calculated and compared with the predefined Z value in24. If the calculated test statistics are larger than 1,645 then the null hypothesis is rejected. Here the null hypothesis corresponds to "There is no similarity between two rankings", where the alternative hypothesis is "Two rankings are similar".

The values of Z between Turkse-Wilson's and Wang's methods and also between Turksen_Wilson's and Biswas's methods are calculated as 4.2540 and 4.1688 correspondingly. These values are greater than 1.645, which indicates that the null hypothesis is rejected and the difference in ranking results is statistically insignificant. From the above discussion, it can be obtained that the ranking of the satisfaction of hotel employees using Turksen-Willson's method are reliable and results can be used by the managers.

Table 5. Maximum similarity degree and ranking using Wang's approach

R Maximum Ranking Linguistic values

similarity degree for each

R1 0.7400 6 very satisfied

R2 0.7384 7 very satisfied

R3 0.7432 6 very satisfied

R4 0.7610 4 very satisfied

R5 0.7755 3 very satisfied

R6 0.7674 4 very satisfied

R7 0.7642 4 very satisfied

R8 0.7900 1 very satisfied

R9 0.7803 2 very satisfied

R10 0.7642 4 very satisfied

R11 0.7690 4 very satisfied

R12 0.7006 8 very satisfied

R13 0.6845 9 very satisfied

R14 0.7368 7 very satisfied

R15 0.7416 6 very satisfied

R16 0.7706 3 very satisfied

R17 0.7642 4 very satisfied

R18 0.7610 4 very satisfied

R19 0.7497 5 very satisfied

20 0.7706 3 very satisfied

Table 6. Maximum similarity degree and ranking using Biswas's approach.

R Maximum similarity degree for each state Ranking Linguistic values

R1 0.3670 7 satisfied

R2 0.3620 7 satisfied

R3 0.3721 6 satisfied

R4 0.3992 4 very satisfied

R5 0.4009 3 very satisfied

R6 0.3911 4 very satisfied

R7 0.3865 5 very satisfied

R8 0.4115 2 satisfied

R9 0.3945 4 satisfied

R10 0.3817 5 satisfied

R11 0.3848 5 very satisfied

R12 0.3322 9 satisfied

R13 0.3217 10 neutral

R14 0.3556 8 satisfied

R15 0.3707 6 satisfied

R16 0.3968 4 satisfied

R17 0.3886 5 satisfied

R18 0.3757 6 satisfied

R19 0.3652 7 very satisfied

R20 0.4211 1 very satisfied

Table 7. Comparisons of Turksen Wilson's, Wang's and Biswas's methods

Traditional Wang's Biswas's Turksen Differences

method method Wilson's method

R Standard Ranking Ranking Ranking Turksen Wilso Turksen Wilson's

Mean Deviation n's & Wang's & Biswas's

R1 3.3427 1.6269 6 7 6 0 -1

R2 3.3427 1.6269 7 7 6 -1 -1

R3 3.4113 1.6466 6 6 5 -1 -1

R4 3.7097 1.7691 4 4 2 -2 -2

R5 3.7379 1.7834 3 3 2 -1 -1

R6 3.7379 1.7834 4 4 3 -1 -1

R7 3.6694 1.7493 4 5 2 -2 -3

R8 3.7702 1.8004 1 2 3 2 1

R9 3.7379 1.7834 2 4 3 1 -1

R10 3.6250 1.7287 4 5 4 0 -1

R11 3.6734 1.7512 4 5 3 -1 -2

R12 3.1935 1.5959 8 9 7 -1 -2

R13 3.0202 1.5813 9 10 8 -1 -2

R14 3.3347 1.6248 7 8 6 -1 -2

R15 3.3790 1.6369 6 6 5 -1 -1

R16 3.6452 1.7379 3 4 4 1 0

R17 3.5403 1.6926 4 5 5 1 0

R18 3.6089 1.7215 4 6 4 0 -2

R19 3.5323 1.6894 5 7 3 -2 -4

R20 3.7944 1.8135 3 1 1 -2 0

r. 0.9759 0.9564

r = 1 —

6Z d2 i=1

n(n2 -1)

Z = rsjn -1

4. Conclusions

In this paper, the integration of fuzzy logic and conjoint analysis is used for evaluation of the job satisfaction of hotel employees. The solution of the problem is based on data collection and analysis of fuzzy preferences of employees. Respondents and experts have stated their opinions about each attribute, and these opinions are used for measuring of job satisfaction. Using preference levels the similarity measure between employees' opinions and experts' opinions are determined. These similarity measures are used for evaluation of job satisfaction of the hotel's employees. The experimental results have been obtained using the opinions of employees of different hotels of North Cyprus. The Turksen-Wilson's method is applied to job satisfaction of the hotel employees. It has been found that 20-th achievement attribute with 9% very dissatisfied, 6% dissatisfied, 11% neutral, 37% satisfied, and 35% very satisfied has the best rank. The states 4, 5 and 7 are ranked as the second. For comparative analysis, the same problem is simulated with Wang's and Biswas's methods also. Using Wang's method the state 8-th state is ranked as first, the state 9 is ranked as second. Using Biswas's method the state 20-th state is ranked as first, and the states 8 and 5 are ranked as second. Based on the simulation results, it can be obtained that the ranking of job satisfaction states using Turksen and Willson's method are reliable and results can be used by the hotel managers. Results obtained from the fuzzy conjoint analysis can be used as an alternative method for analysing job satisfaction.

Appendix A.

The Minnesota Satisfaction Questionnaire

N0 Questions Facets

Item1 Being able to keep busy all the time Activity

Item2 The chance to work alone on the job Independence

Item 3 The chance to do different things from time to ti Variety

Item 4 The chance to be somebody in the community Social Status

Item 5 The way my boss handles his/her workers Supervision / human

relations

Item 6 The competence of my supervisor in making decisions Supervision/ technical

Item 7 Being able to do things that don't go against my conscience Moral values

Item 8 The way my job provides for steady employment Security

Item 9 The chance to do things for other people Social Service

Item 10 The chance to tell people what to do Authority

Item 11 The chance to do something that makes use of my abilities Ability

Item 12 The way company policies are put into practice Policies and practices

Item 13 My pay and the amount of work I do Compensation

Item 14 The chances for advancement on this job Advancement

Item 15 The freedom to use my own judgment Responsibility

Item 16 The chance to try my own methods of doing the job Creativity

Item 17 The working conditions Work Conditions

Item 18 The way my coworkers get along with each other Co- workers

Item 19 The praise I get for doing a good job Recognition

Item 20 The feeling of accomplishment I get from the job Achievement

References

1. Bender K, Donohue S, Heywood J. Job satisfaction and gender segregation. Oxford Economic Papers 2005; vol. 57. No. 3: 479-496.

2. Lacy FJ, Sheehan BA. Job Satisfaction Among Academic Staff: An International Perspective. Higher Education 1997; 34: 305-322.

3. Murray RA. Job satisfaction of Professional and Paraprofessional Library Staff at the University ofNorth Carolina at Chapel Hill,

Unpublished dissertation, University of North Carolina; 1999.

4. Mapesela M, Hay DH. The Effect of Change and Transformation on Academic Staff and Job Satisfaction: A case of a South African

University, Higher Education 2006; 52:711-747.

5. Rajareegam A, Christie Doss I. Analyzing For the Future Expectations on Job Satisfaction of Teachers in Engineering Colleges in

Puducherry. IOSR Journal of Research & Method in Education (IOSR-JRME) 2014; Volume 4(1). Ver. III:46-51.

6. Azadeh A, Rouzbahman M, Saberi M. Utilization of Artificial Intelligence Approach for Assessment of Job Satisfaction. International

Journal of Intelligent Information Technology Application 2009; 2(6): 250-255.

7. Khan AR, Amin HU, Rehman ZU. Application of Expert System with Fuzzy Logic in Teachers Performance Evaluation, International

Journal of Advanced Computer Science and Applications 2011; 2(2): p.51-57.

8. Zadeh LA, Fuzzy Sets. Information and Control 1965; vol. 8:338-353.

9. Arasli H, and Ersoy E. Employability Skills for Hotel Sector : A Comparative Study of Management Expectations and Student Perceptions.

Advances in Business-Related Scientific Research Conference (ABSRC 2010) , Olbia, Sardinia, Italy, September 8-10; 2010.

10. Liu Z, Yang J. A Study on Job Satisfaction of Hotel Employees. 2009 IEEE International Conference on Information Management,

Innovation Management and Industrial Engineering; 2009.

11. Chien MY. Tourism involvement, work engagement and job satisfaction among frontline hotel employees. Annals of Tourism Research

2013; Volume 42:214-239.

12. Lam T, Zhang H, Baum T. An investigation of employees' job satisfaction: the case of hotels in Hong Kong. Tourism Management 2001;

Volume 22. Issue 2: 157-165

13. Rasmani KA, Shahari NA. Job Satisfaction Evaluation using Fuzzy Approach. IEEE Third International Conference on Natural

Computation (ICNC 2007).

14. Lazim MA, Abu Osman MT. Abdullah WSW. Fuzzy Set Conjoint Model in Describing Students' Perceptions on Computer Algebra System

.Learning Environment. IJCSI International Journal oof Computer Science Issues 2011; Vol. 8 Issue 2.

15. Othman M, Ku-Mahamud KR, Abu-Bakar A. Fuzzy evaluation method using fuzzy rule approach in multi criteria analysis. Yugoslav

Journal of Operations Research 2008; 18(1): 95-107.

16. Lazim MA, Abu Osman MT. Measuring teachers' beliefs about Matematics:A fuzzy set approach. World Academy of Science, Engineering

and Technology. International Journal of Social, Management, Economics and Business Engineering 2009; Vol:3(9).

17. Baheri E, Dalvand MR, Ansarinejad A, Miri-Nargesi S, Hatami-Shirkouhi L. A fuzzy conjoint analysis approach for evaluating credit card

services: A case study of Iranian bank. African Journal of Business Management 2011; Vol.5 (7): 2753-2765.

18. Turksen IB, Wilson IA . A Fuzzy Preference Model for Consumer Choice. Fuzzy Sets and Systems 1994; vol. 68(3):253-266.

19. Wang WJ. New similarity measures on fuzzy sets and on elements. Fuzzy Sets Syst 1997; 85(3):305-309.

20. Biswas R. An application of fuzzy sets in students' evaluation. Fuzzy Sets and Systems 1995; 74(2):187-194.

21. YT, and Yurdakul M. Development of a Quick Credibility Scoring Decision Support System Using Fuzzy Topic. Expert Systems with Applications 2010; 37(1).

22. Raju, KS, Kumar DN. Multicriterion Analysis in Engineering and Management. New Delhi: Raj Press; 2010.

23. Srinivasa RK., and Pillai CRS. Multicriterion Decision Making in River Basin Planning and Development. European Journal of Operational

Research 1999; Vol.112(2).

24. Karagul AA, Özdemir A. Evaluation of Financial Information Quality Attributes: A comparison from Turkey. International Journal oof

Business and Social Science 2012; Vol. 3(23).

25. Abiyev RH, Sadikoglu G, Abiyeva E. Fuzzy Evaluation of Job Satisfaction of Hotel Employees. The 2015 World Congress in Computer

Science, Computer Engineering, and Applied Computing, ICAI'15, 27-30 July, Las Vegas, Nevada , USA, 2015