You are on page 1of 7

INTERNATIONAL JOURNAL OF INFORMATION AND SYSTEMS SCIENCES

Volume 1, Number 3-4, Pages 406412

c 2005 Institute for Scientic Computing and Information

APPLYING FUZZY ANALYTIC HIERARCHY PROCESS TO EVALUATE SUCCESS FACTORS OF E-COMMERCE


FENG KONG AND HONGYAN LIU Abstract. : With the fast development of the E-commerce, it becomes critical to set up an E-commerce Evaluation criteria system. Fuzzy Analytic Hierarchy Process is a new multi-criteria evaluation method evolved from Saatys AHP. This paper aimed to nd out the key factors that aect success in E-commerce using fuzzy AHP, and give an evaluation method for E-commerce in order to help researches and managers to determine the drawbacks and opportunities. Key Words. E-commerce website, evaluation, fuzzy AHP.

1. Introduction The Internet has been evolved from a basic tool of communications into a vast and interactive market of products and services involving over 240 million users worldwide [1]. The Internet has the potential to market products and services to customers, to communicate information to a global community, to provide an electronic forum for communications and to process business transactions such as orders and payments. Naturally many enterprises across the world attempt to embrace the digital revolution and place a wide range of materials on the web, from infrastructure to databases to actual services online for the convenience of customers. E-commerce is no longer just an option now but a necessity for enterprises aiming for better performance. Owing to the signicance of E-commerce, the importance of measuring the quality of E-commerce website cannot be overemphasized. With the fast development of the E-commerce, it becomes critical to set up an E-commerce Evaluation criteria system. The purpose of this paper is to develop an E-commerce website assessment method to evaluate the success factors of E-commerce website. The AHP was developed in the 1980s by Saaty [2]. It is a systematic decision making method which includes both qualitative and quantitative techniques. It is being widely used in many elds for a long time. But one of the critical steps of AHP method is to set up the comparison matrixes. When the number of attributes (or alternatives) in the hierarchy increases, more comparisons between attributes (or alternatives) need to be made. This could easily cause confusion due to the excess of questions and hence the eciency of the model. So a consistency check is required for the pair-wise comparison matrix. Therefore, whether the setting of the comparison matrix is scientic aects the correctness of AHP directly. When the comparison matrices are not consistent, we should adjust the elements in the matrixes and carry out a consistency test until they are consistent. In this paper,
Received by the editors June 15, 2004 and, in revised form, January 22, 2005. This research is supported by the Doctorial Teachers Foundation of North China Electric Power University.
406

APPLYING FUZZY AHP TO EVALUATE SUCCESS FACTORS OF E-COMMERCE

407

we introduce a fuzzy AHP in which substitute membership scales for Saatys 1-9 scales to reduce adjusting times needed. This paper is organized as follows: section 2 gives the introduction to the fuzzy analytic hierarchy process; in section 3 we evaluate the success factors of E-commerce website; in section 4 is the conclusion of this paper. 2. The fuzzy Analytic Hierarchy Process 2.1. Denitions of the new fuzzy comparison matrices. The comparison matrix dened by Saaty employs 1-9 scales. The 1-9 scales are illustrated with the following comparison matrix and table 1. (1) A=
w1 w1 w2 w1 wn w1 w1 w2 w2 w2 wn w2

. . .

. . .

w1 wn w2 wn wn wn

. . .

a11 a21 . . . an1

a12 a22 . . . an2

a 1n a 2n . . . ann

Saatys scale 1 3 5 7 9 2, 4, 6, 8

The relative importance of the two sub-elements Equally important Moderately important with one over another Strongly important Very strongly important Extremely important Intermediate values

Table 1. Saatys scale for pairwise comparison. Our new fuzzy comparison matrix diers with Saatys in that we use membership scales, instead of the 1-9 scales, as the values of the elements. (2) A=
w1 w1 +w1 w2 w2 +w1 wn wn +w1 w1 w1 +w2 w2 w2 +w2 wn wn +w2

. . .

. . .

w1 w1 +wn w2 w2 +wn wn wn +wn

. . .

r11 r21 . . . rn1

r12 r22 . . . rn2

r1n r2n . . . rnn

If this comparison matrix is consistent, it should satisfy: 1 1 1 1=( 1) ( 1). rij rik rki

(3)

rii = 0.5, rij + rji = 1,

This method compares weights in pairs and is more straightforward and easier to use for the decision-makers. The meanings of our membership scales can also be expressed in the same way as Saatys scale, see table 2.

408

H. Y. LIU AND F. KONG

Scale values The relative importance of the two sub-elements 0.5 Equally important Slightly important 0.55(or0.5 0.6) 0.65(or0.6 0.7) Important 0.75(or0.7 0.8) Strongly important 0.85(or0.8 0.9) Very strongly important 0.95(or0.9 1.0) Extremely important Table 2. Scale for fuzzy pair-wise comparison. Theoretically, the membership scales put forward in this paper and Saatys scales should satisfy the following: (4) rij = aij . aij + 1

The dierence of our membership scales with Saatys lies in that the values of membership scales falls within the range of [0,1]. 2.2. Calculation of the priority weights. Let (W = w1 , w2 , , wn ), (5) wi = bi
n

. bi

i=1

where, bi =

n j =1

1 n rij

2.3. Consistency test of the comparison matrix. We can use the following equation to calculate the consistency index:
n i=1

(6)

CI =

(AW )i nwi

n1

where the values of the elements in matrix A could be derived by applying equation (3) to matrix R. The comparison matrix will be considered to be consistent if there exists CR = CI RI < 0.1. The various values of RI are shown in table 3. Size of matrix RI 1 0 2 0 3 4 0.58 0.90 5 1.12 6 7 1.24 1.32 8 9 1.41 1.45 10 1.49

Table 3. Values of RI.

APPLYING FUZZY AHP TO EVALUATE SUCCESS FACTORS OF E-COMMERCE

409

3. Evaluation of the success factors of E-commerce website Now we use fuzzy AHP to evaluate the success factors of a commerce website and then undertake a comprehensive evaluation of the website. We will use a numerical illustration to show our method. First, set up the analytic hierarchy model of E-commerce evaluation as the following:

The best success factor

Trust(C1)

System quality(C2)

Content quality(C3)

Online service(C4)

Use (C5)

Security (C11) Privacy(C12)

Visual appearance (C21) response time (C22) page loading speed (C23) 24-hour availability (C24)

Up-to-datedness (C31) Understand ability(C32) timeliness (C33)

Tracking Order Status (C41) Account Maintenance (C42) Payment Alternatives (C43)

Information (C51) Transaction (C52) the disabled service (C53)

Preciseness (C34)

FAQs (C54)

Figure 1. The hierarchy model of the evaluation of websites

Next, we give the fuzzy comparison matrixes of the criteria level and sub-criteria level. For instance, Tables 4-9 show the original fuzzy pair-wise comparison matrixes for E-commerce website evaluation. C1 C2 C3 C4 C5 C1 0.5 0.2 0.3 0.5 0.1 C2 0.8 0.5 0.6 0.9 0.6 C3 0.7 0.4 0.5 0.7 0.3 C4 0.5 0.1 0.3 0.5 0.3 C5 0.9 0.4 0.7 0.7 0.5

Table 4. Fuzzy Comparison matrix at criteria level.

410

H. Y. LIU AND F. KONG

C11 C12

C11 0.5 0.3

C12 0.7 0.5

Table 5. Fuzzy Comparison matrix at sub-criteria level, for criterion C1 . C21 0.5 0.4 0.3 0.6 C22 0.6 0.5 0.4 0.7 C23 0.7 0.6 0.5 0.6 C24 0.4 0.3 0.2 0.5

C21 C22 C23 C24

Table 6. Comparison matrix at sub-criteria level, for criterion C2 .

C31 C32 C33 C34

C31 0.5 0.8 0.3 0.6

C32 0.2 0.5 0.1 0.4

C33 0.7 0.9 0.5 0.8

C34 0.4 0.6 0.2 0.5

Table 7. Comparison matrix at sub-criteria level, for criterion C3 . C41 0.5 0.3 0.4 C42 0.7 0.5 0.6 C43 0.6 0.4 0.5

C41 C42 C43

Table 8. Comparison matrix at sub-criteria level, for criterion C4 .

C51 C52 C53 C54

C51 0.5 0.75 0.3 0.5

C52 C53 0.25 0.7 0.5 0.6 0.4 0.5 0.3 0.6

C54 0.5 0.7 0.4 0.5

Table 9. Comparison matrix at sub-criteria level, for criterion C5

Then calculate the relative priority weights of each criterion and each subcriterion. The results of the instance are shown in table 10.

APPLYING FUZZY AHP TO EVALUATE SUCCESS FACTORS OF E-COMMERCE

411

Criterion C1 C2

Priority of criterion 0.371

0.061

C3

0.153

C4 C5

0.348

0.067

Subcriterion C11 C12 C21 C22 C23 C24 C31 C32 C33 C34 C41 C42 C43 C51 C52 C53 C54

Priority of nal Priority of sub-criterion sub-criterion 0.7 0.260 0.3 0.111 0.278 0.017 0.182 0.011 0.113 0.007 0.427 0.026 0.146 0.022 0.497 0.076 0.062 0.009 0.295 0.045 0.477 0.166 0.207 0.072 0.316 0.110 0.193 0.013 0.432 0.029 0.166 0.011 0.210 0.014

CR of Sub-criterion 0

CR of Criterion

0.049

0.087

0.05

0.095

Table 10. Priority and Consistency ratios We can see that the essential factors aecting the success of websites are factors C1 and C4 ; while C11 is the most critical factor within C1 and C41 is the most critical factor within C4 . Let the evaluation ranking set be: Very good, Good, Moderate, Bad, Very bad. We can give priority ranking weights to the elements of the evaluation set as, respectively, 100, 80, 60, 40, and 20. Suppose the evaluation matrix for the E-commerce website is: Criterion Very Good C11 0.7 C12 0.6 C21 0.2 0 C22 C23 0 C24 0.6 C31 0.6 0.4 C32 C33 0.8 C34 0.7 0.6 C41 C42 0.4 0.3 C43 C51 0 C52 0.2 C53 0 0 C54 Result 0.517 Good 0.2 0.3 0.6 0.3 0.5 0.2 0.1 0.3 0.2 0.2 0.3 0.2 0.4 0.3 0.2 0.2 0 0.263 Moderate Bad 0.1 0 0.1 0 0.2 0 0.4 0.2 0.3 0.2 0.2 0 0.2 0.1 0.1 0.2 0 0 0.1 0 0.1 0 0.3 0.1 0 0.2 0.3 0.3 0.2 0.2 0.6 0.2 0.4 0.1 0.129 0.094 Very Bad 0 0 0 0.1 0 0 0 0 0 0 0 0 0.1 0.1 0.2 0 0.5 0.025

Table 11. Evaluation matrix for websites The comprehensive evaluation result is:

412

H. Y. LIU AND F. KONG

0.517 100 + 0.263 80 + 0.129 60 + 0.094 40 + 0.025 20 = 83.57. Hence, the E-commerce website belongs to grade of good according to our evaluation. 4. Conclusion In this paper, we put forward a subjective and objective integrated approach to determine attributes weights in Fuzzy MADM problems. Since the approach takes into account both subjective and objective factors, it retains the merits of both subjective and objective approaches, to determine weights by solving mathematical models automatically and at the same time take into consideration the decision makers preferences, and overcomes the shortcoming which may happen when either a subjective approach or an objective approach is used in Fuzzy MADM problems. References
[1] Shuojia Guo, Bingjia Shao. Quantitative Evaluation of E-Commercial Websites of Foreign Trade Enterprises in Chongqing. Proceedings of international conference on services systems and services management, 2005: 780-785. [2] T.L. Saaty, The analytic hierarchy process, McGraw-Hill, New York, 1980. [3] N. Bryson, Group decision-making and the analytic hierarchy process: Exploring the consensus-relevant information content, Computer and Operations research, 1996, 23: 2735. [4] J.S. Dyer, A clarication of Remarks on the analytic hierarchy process Management Science, 1990, 36: 274-275. [5] T.L. Saaty, Highlights and critical points in the theory and application of analytic hierarchy process, European Journal of Operational Research, 1997, 74: 426-447. [6] T.L. Saaty, Rank generation, preservation, and reversal in the analytic hierarchy process, Decision Sci., 1987,18: 157-177. [7] E.N. Weiss, V.R. Rao, AHP design issues for largescale systems, Decision Science, 1986, 18: 43-61. [8] Zhang jijun. Fuzzy analytic hierarchy process, A Chinese Journal of Fuzzy systems and mathematics, 1999, 14: 81-89. [9] R. Csutoraa, J.J.Buckleyb. Fuzzy hierarchical analysis: the Lambda-Max method, Fuzzy Sets and Systems, 2001, 120: 181-195. [10] J.J.Buckley, T.Feuring, Y.Hayashi, Fuzzy hierarchical analysis revisited, European Journal of Operational Research, 2001, 129 : 48-64. [11] Fong-Gong Wu, Ying-Jye Lee, Ming-Chyuan Lin, Using the fuzzy analytic hierarchy process on optimum spatial allocation, International Journal of Industrial Ergonomics, 2004,33: 553569. [12] Li Rongjun, The theory and its applications of Fuzzy MCDM, Science Press, Beijing, 2000. [13] R. M. Tong, P.P Bonissone, A linguistic approach to decision making with fuzzy sets, IEEE Trans. System Man Cybernet, Vol 10, pp. 10,716-723, 1980. [14] Carlsson C, Fuller R., Fuzzy multiple criteria decision making: recent developments, Fuzzy sets and Systems, Vol 78, 139-153, 1996. [15] M. Delgado, F. Herrera, et al. Combining numerical and linguistic information in group decision making. Journal of information sciences, 107(1998): 177-194. Economics and Management Department, North China Electric Power University, Baoding, 071003, China. E-mail : k-xx1997@163.com and wenfeng3596@sina.com

You might also like