E-Commerce Assessment in Fuzzy Situation 21

Size: px
Start display at page:

Download "E-Commerce Assessment in Fuzzy Situation 21"

Transcription

1 E-Commerce Assessment n Fuzzy Stuaton 21 X 2 E-Commerce Assessment n Fuzzy Stuaton Mehd Fasanghar Iran Telecommuncaton Research Center (ITRC) 1. Introducton Customer satsfacton degree s crtcal for e-commerce enterprses. Thus, more and more commercal organzatons attend to customer satsfacton as ther man strategy (Mhels et al., 2001). Customer satsfacton means the satsfacton degrees of customers purchasng commodtes. Under electronc commerce, how to rase the consumers degree of satsfacton and gan the consumers loyalty has become the key factor relatng wth whether ecommerce enterprse can survve. The vew of the phlosophy of modern management scentfc holds that, "customer satsfacton s the basc crtera of enterprse". Nowadays, more and more commercal organzatons take customer satsfacton as ther man strategy obect (Mhels et al., 2001). Customer satsfacton and customer satsfacton ndex have been wdely developed n both theory and applcatons (C. Albert, 2002, E.W. Anderson & C. Fornell, 2000, P. Hackl et al., 2000, D.J. Mchael, 2004). Fornell (C. Fornell, 1992) developed the frst model of customer satsfacton n The Amercan customer satsfacton ndex (ASCI) was set up n 1994 (Yn Rongwu, 2000), Swedsh SCSI, European ECSI and Korean KCSI etc. Chnese Customer satsfacton Index (CCSI) started n 1998, are stll on the stage of exploraton and learnng (Zhao Pengxang, 2001). Customer satsfacton has been used frequently n the qualty system certcaton of ISO9000 of 2000 edton, whch shows the mportant of ths concept (Xaohong Lu et al., 2008). Also, Many countres are conductng Customer Satsfacton Index studes snce Customer Satsfacton Index can be used as a predcator for proftablty and value of the organzatons (S.-H. Hsu, 2008). Nowadays, more attenton has been pad to the problem of e-commerce customer satsfacton (M. Wang, 2003). The e-commerce customer satsfacton parameters should be vsble for customers. The more mportant parameters have been chosen based on the lterature revew: ndependency, comparable, and feasblty (Mhels et al., 2001, Yan Xao- Tan & We Hong-Jun, 2005, Yn Rongwu, 2000, Zheng Yue-Fang, 2005, Lu, 2007). Each ndex should ndependently represent the servce qualty satsfacton from some aspect (Fasanghar & Roudsar, 2008). Whle customers represent ther satsfacton value of e- commerce, the ndexes should be comparable for dfferent customer. Thus, the ndexes of e- commerce customer satsfacton have been obtaned based on lterature revew so that support all of the customer satsfacton area for e-commerce.

2 22 E-Commerce To evaluate the e-commerce customer satsfacton quanttatvely, many countres establshed ther own ndex of customer satsfacton degree, namely customer satsfactory Index, whch s a new set of ndexes evaluatng an enterprse and a trade or an ndustry completely from customer s angle (Lu Pede, 2007). The e-commerce customer satsfacton should pay attenton to the parameters whch are vsble for customers. The more mportant parameters have been chosen based on the lterature revew: ndependency, comparable, and feasblty (Mhels et al., 2001, Yan Xao- Tan & We Hong-Jun, 2005, Yn Rongwu, 2000, Zheng Yue-Fang, 2005). Customer evaluaton for e-commerce would be possble f the ndependency among the ndexes exst. The selecton of ndexes, therefore, should be hgh enough n resoluton to help dstngush the factors. Each ndex should ndependently represent the servce qualty satsfacton from some aspect. To clarfy, ndexes should be comparable, as the model should evaluate the dfferent customer nputs, whch express the satsfacton value for e-commerce; consequently, the ndexes should be comparable for dfferent customer whle they represent ther satsfacton value of e-commerce. At last, dentfcaton and reducton of customer satsfacton are the obectve of e-commerce customer satsfacton. The ttle and content of each ndex, hence, should be well understood by the customers. In ths paper, n the next secton, fuzzy set theory and the prncpals of trangular fuzzy number have been presented. Then, the model for e-commerce customer satsfacton has been llustrated. A case study has been done, as valdaton of presented method; fnally, concluson has been presented. 2. Fuzzy Set Theory Fuzzy set theory provdes a framework for handlng the uncertantes. Zadeh ntated the fuzzy set theory (Zadeh L. A., 1965). Bellman presented some applcatons of fuzzy theores to the varous decson-makng processes n a fuzzy envronment (Bellman R. E. & Zadeh L. A., 1970). In non-fuzzy set every obect s ether a member of the set or t s not a member of the set but n fuzzy sets every obect s to some extent member of a set and to some extent t s member of another set. Thus, unlke the crsp sets membershp s a contnuous concept n fuzzy sets. Fuzzy s used n support of lngustc varables and there s uncertanness n the problem. Fuzzy theory s wdely applcable n nformaton gatherng, modelng, analyss, optmzaton, control, decson makng, and supervson. Specal cases of fuzzy numbers nclude crsp real number and ntervals of real numbers. Although there are many shapes of fuzzy numbers, the trangular and trapezodal shapes are used most often for representng fuzzy numbers. The followng descrbes and defntons show that membershp functon of trangular fuzzy number, trapezodal fuzzy number, and ts operatons. A fuzzy number à s convex, f [ x (1 ) x ] mn[ ( x ), ( x )]. x, x X, [0,1] A 1 2 A 1 A (1) Alternatvely, a fuzzy set s convex f all α-level sets are convex.

3 E-Commerce Assessment n Fuzzy Stuaton 23 A fuzzy set à n the unverse of dscourse X s normal f (A. Kaufmann & M.M. Gupta, 1988, S. Mabuch, 1988) sup ( ) 1 (2) x A x A nonempty fuzzy set à can always be normalzed by ( x ) / sup ( x ) A x A. A trangular fuzzy number can be defned by A ( a1, a2, a3), where a a a, ts member functon represented as follows. 0 x a1 ( x a1) a ( 1 x a2 a2 a1) A (3) ( x a3) a ( 2 x a a 3 3 a2) 0 x a3 Let A and B be two fuzzy numbers parameterzed by the ( a1, a2, a 3) and ( b 1, b 2, b 3 ), respectvely. Then the operatons of trangular fuzzy numbers are expressed as (S.J. Chen & C.L. Hwang, 1992a): A ( ) B ( a1, a2, a3) ( b1, b2, b3) ( a1 b1, a2 b2, a3 b3) A ( ) B ( a1, a2, a3) ( b1, b2, b3) ( a1 b1, a2 b2, a3 b3) (4) A ( ) B ( a1, a2, a3) ( b1, b2, b3) ( a1 b1, a2 b2, a3 b3) A ( ) B ( a, a, a ) ( b, b, b ) ( a b, a b, a b ) Trangular fuzzy numbers are approprate for quantfyng the vague nformaton about most decson problems (C.H. Cheng & Y. Ln, 2002), and the prmary reason for usng trangular fuzzy numbers can be stated as ther ntutve and computatonal-effcent representaton. In ths paper, the trangular fuzzy number s used for measurng customer satsfacton. More detals about arthmetc operatons laws of trapezodal fuzzy number can be seen n (Lee et al., 2004). Consderng experts E provde the satsfacton degree wth A ( ). The evaluaton values gven by each expert ( E ) are presented n the form of a trangular fuzzy number: A ( a, a, a ), where 1,2,..., n (5) ( ) ( ) ( ) ( ) The average A of all A m ( ) s computed usng average means A ( a, a, a ) ( a, a, a ) (6) n n n ( ) ( ) ( ) m m1 m 2 m n 1 n 1 n 1 For defuzzfcaton of A m, the followng formula can be used:

4 24 E-Commerce A a 2a a 4 m1 m 2 m 3 m (7) 3. The Model for E-Commerce Customer Satsfacton Evaluaton 3.1 Determnaton of Indexes Some of the researches constructed evaluaton ndex from commercal content, customer s concern, effectve navgaton, webste desgn, safety, convenence, merchandse plannng, contact convenence, customer servce nformaton, convenence of gettng product nformaton, accuracy, content relatedness, ntegrty, varety n dsplayng, nformaton tmely updatng, easy applcaton, system rapdty, servce response n tme, and guaranteed servce (Duo Q, 2003, Lu Pede, 2007, Yu Hongyan, 2006, Gao Dan, 2004, Gan Yong, 2006). Based on lterature, ths paper constructs customer satsfacton ndex of BtoC e-commerce enterprse, and evaluates customer satsfacton of BtoC e-commerce enterprse by adoptng Fuzzy Trangular Numbers for Lngustc Varables and usng fuzzy TOPSIS method Table1. Obect Indcator Descrpton Customzaton (Pr1) The degree of customer needs satsfyng Value (Pr 2) The unque product characterstcs Product Informaton (P r3) Accessblty to the nformaton of the products Scope (Pr 4) The dmensons whch the product satsfy them Accuracy of qualty (Pr 5) The qualty parameters satsfacton Guaranty (Pr 6) The confdence of the product qualty Atttude (S1) How servces are receved Informaton (S2) Accessblty to the nformaton of the servces Dstrbuton (S3) The method and tools of servce delvery Servce Response and feedback (S4) The qualty of feedback of servces Call center (S5) Avalablty of a call center for customer care Qualty (S6) The qualty parameters satsfacton Management (S7) Determnaton of servce process and control ndexes Safety (N1) Securty degree of networks Relablty (N2) The amounts of rsk Network Operablty (N3) The network support of the customer needs system Accessblty (N4) The avalablty of network n 24*7 Humanzaton (N5) Consderaton of human computer nteracton Accuracy of fee calculaton (Pa1) The trust n fnancal computaton payment Accuracy of fee collecton (Pa2) The trust to the system of fee collecton Method (Pa3) The techncal methods relatng to the payment Facltes (Pa4) The qualty and sutablty of facltes used for payment Table 1. Indexes of e-commerce customer satsfacton evaluaton 3.2 Determnaton the Weght of Indexes TOPSIS (Technque for Order Preference by Smlarty to an Ideal Soluton) method s presented n Chen and Hwang (S.J. Chen & C.L. Hwang, 1992b), wth reference to Hwang and Yoon (C.L. Hwang & K. Yoon, 1981). TOPSIS s a multple crtera method to dentfy solutons from a fnte set of alternatves. The basc prncple s that the chosen alternatve should have the shortest dstance from the postve deal soluton and the farthest dstance

5 E-Commerce Assessment n Fuzzy Stuaton 25 from the negatve deal soluton. The procedure of fuzzy TOPSIS can be expressed n a seres of steps: E-commerce customer satsfacton s a typcal Mult Crtera Decson Makng (MCDM) problem. An MCDM problem can be formulated as follows: C... C... C X k 1 A x... x... x 1 11k 1 k 1nk A x... x... x 1k k nk A x... x... x m m1k mk mnk w 1k W k w k (9) w nk where K 1,..., p decson makers select an alternatve out of 1,..., m alternatves ( A ) accordng to 1,..., n crtera ( C ). x k s the performance of alternatve A wth respect to crteron C estmated by decson maker k, whle W k s the weght of crteron C gven by the decson maker k. n (8) X k and W k are respectvely the decson matrx and the weght vector for the decson maker k. In the fuzzy TOPSIS procedure, the crtera weghts ( w, 1, 2,... trangular fuzzy number of crtera) and characterstc values of alternatves at crtera ( x, 1,2,... number of alternatves, 1,2,... number of crtera) are nputs and placed n matrx form (P. Sen & J-B. Yang, 1998, C.-T. Chen, 2000, Fasanghar et al., 2008) as shown n step 1. Step 1: Inputs are expressed n matrx format as; x 11 x x 1n x 21 x x 2n (10) x m1 x m 2... x mn W w 1, w 2,..., w n (11) w are trangular fuzzy numbers as w ( w 1, w 2, w 3). Step 2: the normalzed decson matrx s constructed usng equaton (12) (H.S. Byun & K.H. Lee, 2004, P. Sen & J-B. Yang, 1998). n x n 1 x 2 (12) Step 3: The weghted normalzed decson matrx s,

6 26 E-Commerce where, V v, 1,2,..., m, 1,2,..., n. (13) m n v n w (14) Set 4: each fuzzy number s defuzzfed usng equaton(15). For trangular fuzzy number v ( a, b, c ) ts defuzzfcaton value s defned to be And defuzzfed weghted normalzed matrx determned as 1,2,..., n. v a 2b c (15) 4 V v, 1,2,..., m, Step 5: The deal soluton, A ( A ; 1,2,..., N ), s made of all the best performance scores and the negatve-deal soluton, A ( A ; 1,2,..., N ), s made of all the worst performance scores at the measures n the defuzzfed weghted normalzed decson matrx. They ate calculated usng equatons (16) and (17). In these equatons, the measures can be dvded nto two classes: the frst s of an nput or cost nature, so that smaller performance scores at these measure are preferred; the second s of an output or beneft nature and larger performance scores at these measures are preferred (H.S. Byun & K.H. Lee, 2004, P. Sen & J- B. Yang, 1998). 1 2 n 1 2 n A v, v,..., v (max v I ),(mn v J ), (16) A v, v,..., v (mn v J ),(max v I ). (17) m n Step 6: The dstance of an alternatve to the deal soluton ( d ), and from the negatve deal soluton ( d ) are calculated usng equaton (18) and (19) (H.S. Byun & K.H. Lee, 2004, P. Sen & J-B. Yang, 1998). m 2 ( ) 1,2,..., ; 1,2,...,, 1 d v v n m (18) m 2 ( ) 1,2,..., ; 1,2,..., 1 d v v n m (19) Step 7: The rankng score ( R ) s calculated usng equaton (20) (H.S. Byun & K.H. Lee, 2004, P. Sen & J-B. Yang, 1998). d R, 1,2,.., m. d d (20)

7 E-Commerce Assessment n Fuzzy Stuaton 27 The obect of fuzzy TOPSIS method s to choose the alternatves that have the shortest dstance from the postve deal soluton and the farthest dstance from the negatve deal soluton. Even though TOPSIS ntroduce two reference pont, t does not consder the relatve mportance of the dstance from the ponts (G.R. Jahanshahloo et al., 2006). 4. Case study In order to assess the proposed methodology, we select 5 e-commerce webstes as our alternatves and 10 customers whch procure ther needful. Experts fll n the questonnare wth fuzzy trangular numbers. Table 1 presents the ndex of the e-commerce assessment; we use these ndexes for the selected webstes assessment. As the frst step, equaton (10) s formed as Table 2. The weghts of the ndexes are equal n vew pont of the experts. Steps 2, 3, 4 and 5 are done and the defuzzfed weghted normalzed decson matrx, A, and A s computed. In step 6, the d 1, d 2, d 3, d 4, d 5, d 1, d 2, d 3, d 4,andd are calculated respectvely: , , , , , , , , , and The fnal results of fuzzy TOPSIS method s obtaned as R 1=0.6339, R 2 =0.6602, R , R , and R Obvously, the thrd webste (alternatve number 3) s n frst row of the webste rankng and has the hghest score n rankng. Unlke, the ffth webste (alternatve number 5) has the lowest score n rankng method of fuzzy TOPSIS. 5. Conclusons The man contrbuton of ths paper s proposng a rankng method for assessng the e- commerce under uncertan stuatons. In fact, combnaton of fuzzy trangular rubbers, TOPSIS method, and e-commerce ndexes s proposed n ths paper. Hence we can assess the customer satsfacton of e-commerce, and we run a case study n whch the 5 e- commerce webstes are assessed wth 10 experts of e-commerce who are famlar wth the selected webstes. Fortunately, all of the experts are pleased of the obtaned results.

8 28 E-Commerce Index Fuzzy number Alternatves Index Fuzzy number Alternatves Index Fuzzy number Alternatves A1 A2 A3 A4 A5 A1 A2 A3 A4 A5 A1 A2 A3 A4 A5 a a a Pr1 a S3 a N4 a a a a a a a Pr2 a S4 a N5 a a a a a a a Pr3 a S5 a Pa1 a a a a a a a Pr4 a S6 a Pa2 a a a a a a a Pr5 a S7 a Pa3 a a a a a a a Pr6 a N1 a Pa4 a a a a a a S1 a N2 a a a a a S2 a N3 a a a Table 2. E-commerce customer satsfacton value for the 5 selected e-commerce webste 6. References A. Kaufmann & M.M. Gupta (1988). Fuzzy Mathematcal Models n Engneerng and Management Scence, North-Holland, Amsterdam Bellman R. E. & Zadeh L. A. (1970). Decson makng n a fuzzy envronment. Management Scences, Vol. 17, C.-T. Chen (2000). Extensons of the Topss for group decson-makng under fuzzy envronment. Fuzzy Sets and Slstems, Vol. 114, 1-9 C. Albert (2002). Servce Loyalty: the effects of servce qualty and the medatng role of customer satsfacton. Eur. J. Market., Vol. 36, C. Fornell (1992). A natonal customer satsfacton barometer: the swedsh experence. J. Market., Vol. 56, 6-21 C.H. Cheng & Y. Ln (2002). Evaluatng the best man battle tank usng fuzzy decson theory wth lngustc crtera evaluaton. European ournal of operaton research, Vol. 142, No. 2,

9 E-Commerce Assessment n Fuzzy Stuaton 29 C.L. Hwang & K. Yoon (1981) Multple Attrbute Decson Makng Methods and Applcatons. Berln Hedelberg, Sprnger. D.J. Mchael (2004). Determnng attrbute mportance n a servce satsfacton model. J. Serv. Res., Vol. 7, Duo Q (2003). Analyse and desgn on customer satsfactory system under E-commerce. Sc- Technology and Management, Vol. 1 E.W. Anderson & C. Fornell (2000). Foundatons of the Amercan Customer Satsfacton Index. Total Qual. Manag, Vol Fasanghar, M.; Gholamy, N.; Chaharsoogh, S. K.; Qadam, S. & Delgosha, M. S. (2008). The Fuzzy Evaluaton of E-Commerce Customer Satsfacton Utlzng Fuzzy TOPSIS, Proceedng of Internatonal Symposum on Electronc Commerce and Securty (ISECS 2008), pp , Chna, IEEE Computer Socety, Chna Fasanghar, M. & Roudsar, F. H. (2008). The Fuzzy Evaluaton of E-Commerce Customer Satsfacton. World Appled Scences Journal, Vol. 4, No. 2, G.R. Jahanshahloo; F. Hossenzadeh Lotf & M. Izadkhah (2006). An algorthm method to extend TOPSIS for decson-makng problems wth nterval data. Appled Mathematcs And Computaton, Vol. 175, Gan Yong (2006) Research on the Fuzzy Comprehensve Evaluaton of Customer Satsfacton n B2C Electronc Busness Enterprse. Maseter dssertaton of Jln Unversty. Gao Dan (2004) Smple analyse on evaluaton ndcator system of Custom Satsfacton n E- commerce. Chna ECommerce. H.S. Byun & K.H. Lee (2004). A decson support system for the selecton of a rapd prototypng process usng the modfed TOPSIS method. Internatonal Journal of Advanced Manufacturng Technology, 1-10 Lee, J. W.; Hong, E. & Park, J. (2004). A Q-learnng based approach to desgn of ntellgent stock tradng agents. Engneerng Management Conference, Proceedngs. IEEE Internatonal, Vol. 3, Lu, P. (2007). Evaluaton Model of customer Satsfacton B2C E-Commercebased on Combnaton of Lngustc Varables And Fuzzy Trangular Numbers, Proceedng of eght ACIS nternatonal conference on software engneerng, artfcal ntellgence, networkng, and parallel/ dstrbuted computng, pp Lu Pede (2007). Evaluaton Model of Customer Satsfacton of B2C E_Commerce Based on Combnaton of Lngustc Varables and Fuzzy Trangular Numbers, Proceedng of Eghth ACIS Internatonal Conference, pp , Qngdao, Chna, Qngdao, Chna M. Wang (2003). Assessment of e-servce qualty va e-satsfacton n e-commerce globalzaton. The Electr. J. Inform. Syst. Develop. Countres Vol. 11, No. 10, 1-4 Mhels, G.; Grgorouds, E. & Sskos, Y. (2001). Customer satsfacton measurement n the prvate bank secton. European Journal of Operaton Research, Vol. 130, P. Hackl; K. Krstensen & A.H. Westlund (Eds.) (2000). Customer Satsfacton: Theory and Measurement. JSpecal Issue of ournal of Total Qualty Management P. Sen & J-B. Yang (1998). Multple Crtera Decson Support n Engneerng Desgn, Sprnger Verlag London Lmted, London, Great Brtan S.-H. Hsu (2008). Developng an ndex for onlne customer satsfacton: Adaptaton of Amercan Customer Satsfacton Index. Expert Systems wth Applcatons, Vol. 34,

10 30 E-Commerce S. Mabuch (1988). An approach to the comparson of fuzzy subsets wth an α-cut dependent ndex. IEEE Transactons on Systems, Man, and Cybernetcs SMC, Vol. 18, No. 2, S.J. Chen & C.L. Hwang (1992a). Fuzzy Multple Attrbute Decson Makng: Methods and Applcaton, Sprnger, New York S.J. Chen & C.L. Hwang (1992b) Fuzzy Multple Attrbute DecsonMakng:Methods and Applcatons. Berln, Sprnger-Verlag. Xaohong Lu; Xany Zeng; Yang Xu & Ludovc Koehl (2008). A fuzzy model of customer satsfacton ndex n e-commerce. Mathematcs and Computers n Smulaton, Vol Yan Xao-Tan & We Hong-Jun (2005). Constructon and applcaton of power customer satsfacton degree evaluaton system. East Chna Electrc Power, Vol. 33, No. 12, Yn Rongwu (2000). revew of customer satsfactory Index n US. World Standardzaton & Qualty Management, Vol. 1, No. 1, 7-10 Yu Hongyan (2006). Bref Analyss on Custom Satsfacton BtoC n E-commerce. Journal of Hunan Unversty of Scence and Engneerng, Vol. 1 Zadeh L. A. (1965). Fuzzy sets. Informaton and Control, Vol. 8, Zhao Pengxang (2001). Research on Buldng and Performance of Customer Satsfacton Management System. World Standardzaton & Qualty Management, Vol. 6, No. 6, Zheng Yue-Fang (2005). Customer satsfacton-the most mportant task of electrcty servce. Chna Qualty, Vol. 3, No

11 E-commerce Edted by Kyeong Kang ISBN Hard cover, 284 pages Publsher InTech Publshed onlne 01, February, 2010 Publshed n prnt edton February, 2010 E-commerce provdes mmense capablty for connectvty through buyng and sellng actvtes all over the world. Durng the last two decades new concepts of busness have evolved due to popularty of the Internet, provdng new busness opportuntes for commercal organsatons and they are beng further nfluenced by user actvtes of newer applcatons of the Internet. Busness transactons are made possble through a combnaton of secure data processng, networkng technologes and nteractvty functons. Busness models are also subected to contnuous external forces of technologcal evoluton, nnovatve solutons derved through competton, creaton of legal boundares through legslaton and socal change. The man purpose of ths book s to provde the reader wth a famlarty of the web based e- commerce envronment and poston them to deal confdently wth a compettve global busness envronment. The book contans a numbers of case studes provdng the reader wth dfferent perspectves n nterface desgn, technology usage, qualty measurement and performance aspects of developng web-based e-commerce. How to reference In order to correctly reference ths scholarly work, feel free to copy and paste the followng: Mehd Fasanghar (2010). E-Commerce Assessment n Fuzzy Stuaton, E-commerce, Kyeong Kang (Ed.), ISBN: , InTech, Avalable from: InTech Europe Unversty Campus STeP R Slavka Krautzeka 83/A Reka, Croata Phone: +385 (51) Fax: +385 (51) InTech Chna Unt 405, Offce Block, Hotel Equatoral Shangha No.65, Yan An Road (West), Shangha, , Chna Phone: Fax: