Data Analysis Machine Learning and Applications Episode 2 Part 8 docx

In this paper, we compare a new online version of AHP with Adaptive Conjoint AnalysisACA on the basis of a comprehensive empirical study in tourism which includes 10 attributesand 35 att

Fig 5 U*-Map (Island View)

Fig 6 U*Matrix and Result of U*-C-Algorithm

5 Conclusion

The authors present a classification approach in connection with geospatial data Thecentral issue of the grouping processes are the shrinking and growing phenomena inGermany First the authors examine the pool of data and show the importance for theinvestigation of distributions according to the dichotomic properties Afterwards it isshown that the use of Emergent SOMs is an appropriate method for clustering and

Fig 7 Localisation of Shrinking and Growing Municipalities in Germany

classification The advantage is to visualize the structure of data and later on to define

a number of feasible cluster using U*C-algorithm or manual bestmatch grouping cesses The application of existing visual methods especially U*-Matrix shows that it

pro-is possible to detect meaningful classes among a large amount of geospatial objects.For example typical hierarchical algorithm would fail to examine 12430 objects Assuch, the authors believe that the presented procedure of the wise classification andthe ESOM approach complements the former proposals for city classification It isexpected that in the future the concept of data mining in connection with knowledgediscovery techniques will get an increasing importance for the urban research andplanning processes (Streich, 2005) Such approaches might lead to a benchmark sys-tem for regional policy or other strategical institutions To get more data for a deeperempirical examination it is necessary to conduct field investigation in selected areas

BEHNISCH, M (2007): Urban Data Mining Doctoral thesis, Karlsruhe (TH).

DEMSAR, U (2006): Data Mining of geospatial data: combining visual and automatic ods Urban Planning Department, KTH Stockholm.

meth-HAND, D and MANNILA, H (2001): Principles of Data Mining MIT Press.

KASKI et al (1999): Analysis and Visualization of Gene Expression Data using Self ing Maps, Proc NSIP.

Organiz-KOHONEN, T (1982): Self-Organizing formation of topologically correct feature maps

Bi-ological Cybernetics 43, 59-69

RIPLEY, B (1996): Pattern Recognition and Neural Networks Cambridge Press.

STREICH, B (2005, S 193 ff.): Stadtplanung in der Wissensgesellschaft - Ein Handbuch VS

Verlag für Sozialwissenschaften, Wiesbaden

ULTSCH, A (1999): Data Mining and Knowledge Discovery with Emergent Self Organizing Feature Maps for Multivariate Time Series, In: Oja, E., Kaski, S (Eds.): Kohonen Maps,

Classifi-ULTSCH, A (2005): U*C Self-organized Clustering with Emergent Feature Map, In

Prooceedings Lernen, Wissensentdeckung und Adaptivität (LWA/FGML 2005), brücken, Germany, pp 240-246

Saar-ULTSCH, A (2007): Mining for Understandable Knowledge Submitted.

Martin Meißner, Sören W Scholz and Reinhold Decker

Department of Business Administration and Economics, Bielefeld University,

33501 Bielefeld, Germany

{mmeissner, sscholz, rdecker}@wiwi.uni-bielefeld.de

Abstract The Analytic Hierarchy Process (AHP) has been of substantial impact in business

research and particularly in managerial decision making for a long time Although empiricalinvestigations (e.g Scholl et al (2005)) and simulation studies (e.g Scholz et al (2006)) haveshown its general potential in consumer preference measurement, AHP is still rather unpopular

in marketing research

In this paper, we compare a new online version of AHP with Adaptive Conjoint Analysis(ACA) on the basis of a comprehensive empirical study in tourism which includes 10 attributesand 35 attribute levels We particularly focus on the convergent and the predictive validity

of AHP and ACA Though both methods clearly differ regarding their basic conception, theresulting preference structures prove to be similar on the aggregate level On the individuallevel, however, the AHP approach results in a significantly higher accuracy with respect tochoice prediction

1 Preference measurement for complex products

Conjoint Analysis (CA) is one of the most prominent tools in consumer preferencemeasurement and widely used in marketing practice However, an often stated prob-lem of full-profile CA is that of dealing with large numbers of attributes This limi-tation is of great practical relevance because ideally all attributes and attribute levelsthat affect individual choice should be included to map a realistic choice process.Various methods have been suggested to provide more accurate insights into con-sumer preferences for complex products with many attributes (Green and Srinivasan(1990)) Self-Explicated (SE) approaches, e.g., are used to minimize the informationoverload by questioning the respondents about each attribute separately But SE hasbeen criticized for lacking the trade-off perspective underlying CA For this reason,hybrid methods combining the strengths of SE and full-profile CA have been de-veloped Sawtooth Software’s ACA is a commercially successful computer-basedtool facilitating efficient preference measurements for complex products (for de-tails, please see Sawtooth Software (2003)) While several other approaches, such asthe hierarchical Bayes extensions of Choice-Based Conjoint Analysis, are availablefor estimating part-worth utilities on the individual level, ACA is still the standard

in preference measurement for products with more than six attributes (Hauser andToubia (2002), Herrmann et al (2005)) and widely used in marketing practice (Saw-tooth Software (2005)) In this paper, ACA will set a common benchmark for ourempirical comparison.

Against this background, we introduce an online version of AHP as an tive tool for consumer preference measurement in respective settings Initially, AHPhas been developed to analyze complex decision problems by decomposing themhierarchically into better manageable sub-problems It has been of substantial im-pact in business research and particularly in managerial decision making for a longtime Empirical investigations (e.g Scholl et al (2005)) and simulation studies (e.g.Scholz et al (2006)) recently demonstrated its general potential in consumer pref-erence measurement However, to the best of our knowledge, AHP has never beentested in a real-world online consumer survey, even though internet-based survey-ing gains increasing importance (Fricker et al (2005)) In this paper, we compare anonline version of AHP with ACA by referring to a comprehensive empirical investi-gation in tourism which includes 10 attributes and 35 attribute levels

alterna-The remainder of the paper is structured as follows: In Section 2, we brieflyoutline the methodological basis of AHP Section 3 describes the design of the em-pirical study The results are presented in Section 4 and we conclude with some finalremarks in Section 5

2 The Analytic Hierarchy Process – AHP

In AHP, a decision problem, e.g determining the individually most preferred tive from a given set of products, has to be arranged in a hierarchy It is referred to asthe “main goal" in the following and represented by the top level of the hierarchy Bydecomposing the main goal into several sub-problems, each of them representing therelation of a second level attribute category with the main goal, the complexity of theoverall decision problem is reduced The individual attribute categories, on their part,are broken down into attributes and attribute levels defining “lower" sub-problems.Typically, different alternatives (here: products or concepts) are considered at thebottom level of the hierarchy But due to the large number of hypothetical products,

alterna-or rather “stimuli" in the CA terminology, the use of incomplete hierarchies onlycovering attribute levels, instead of complete stimuli at the bottom level, is advis-able

For the evaluation of summer vacation packages–the objects of investigation inour empirical study–we have structured the decision problem in a 4-level hierar-chy The hierarchical structure displayed in Table 1 reflects the respondents’ averageperceptions and decomposes the complex product evaluation problem into easy toconceive sub-problems

First, the respondents have to judge all pairs of attribute levels of each problem on the bottom level of the hierarchy Then, they proceed with paired compar-isons on the next higher level of the hierarchy, an so on In this way, the respondentsare first introduced to the attributes’ range and levels

sub-Table 1 Hierarchical structuring of the vacation package evaluation problem Attribute Attribute Attribute levels

category

Vacation spot Sightseeing offers 1) Many 2) Some 3) Few

Security concerns 1) Very high 2) High 3) AverageClimate 1) Subtropical 2) Mediterranean 3) DesertBeach 1) Lava sand 2) Sea sand 3) Shingle

Hotel Leisure 1) Fitness room 2) Lawn sport facilities

services activities 3) Aquatic sports facilities

4) Indoor swimming pool 5) Sauna6) Massage parlor

Furnishing 1) Air conditioning 2) In-room safe

3) Cable/satellite TV 4) BalconyCatering 1) Self-catering 2) Breakfast only

3) Half board 4) Full board 5) All-inclusive

Hotel facilities Location 1) Near beach 2) Near town

Type of building 1) Rooming house 2) Hotel complex

3) BungalowOutside facilities 1) Several pools 2) One large pool

3) One small pool

In order to completely evaluate a sub-problem h with n helements,n h (n h −1)

2 wise comparisons have to be carried out Intuitively, the hierarchically decomposition

pair-of complex decision problems in many small sub-problems reduces the number pair-ofpaired comparisons that have to be conducted to evaluate the decision problem.Each respondent has to provide two responses for each paired comparison First,

the respondent has to state the direction of his or her preference for element i pared to element j with respect to an element h belonging to the next higher level.

com-Second, the strength of his or her preference is measured on a 9-point ratio-scale,

where 1 means “element i and j are equal" and 9 means “element i is absolutely ferred to element j" (or vice versa) The respondent’s verbal expressions are trans- formed into priority ratios a h

pre-i j, where a large ratio expresses a distinct preference

of i over j in sub-problem h The reciprocal value a h = 1/a h

i j indicates the

prefer-ence of element j over i All pairwise comparisons of one sub-problem measured

with respect to a higher level element h are brought together in the matrix A h(Saaty(1980)):

Starting from these priority ratios a h

i j , the relative utility values w h

i are calculated by

solving the following eigenvalue problem for each sub-problem h:

Ahwh= Oh

The normalized principal right eigenvector belonging to the largest eigenvalue Oh

max

of matrix Ahyields the vector wh , which contains the relative utility values w h

i for

each element of sub-problem h.

An appealing feature of AHP is the computability of a consistency index (CI),which describes the degree of consistency in the pairwise comparisons of a con-

sidered sub-problem h The CI value expresses the relative deviation of the largest

eigenvalue Oh

maxof matrix Ah from the number of included elements n h:

CI h=Oh max − n h

To get a notion of the consistency of matrix Ah , CI his related to the average

consis-tency index of random matrices RI of the same size The resulting measure is called the consistency ratio CR h , with CR h=CI h

RI In order to evaluate the degree of

consis-tency for the entire hierarchy, the arithmetic mean of all consisconsis-tency ratios ACR can

be used (Saaty (1980))

The AHP hierarchy can be represented by an additive model according to attribute value theory In doing so, the part-worth utilities are determined by multi-plying the relative utility values of each sub-problem along the path, from the maingoal to the respective attribute level The attribute importances are calculated by mul-tiplying the relative utility values of the attribute categories with the relative utilityvalues of the attributes with respect to the related category Then, the overall utility

multi-of a product or concept stimulus is derived by summing up the part-worth utilities multi-ofall attribute levels characterizing this alternative

3 Design of the empirical study

The attributes and levels considered in the following empirical study were mined by means of dual questioning technique Repertory grid and laddering tech-niques were applied to construct an average hierarchically representation of the prod-uct evaluation problem (Scholz and Decker (2007)) Altogether, 200 respondents par-ticipated in these pre-studies The resulting product description design (see Table 1)was used for both the AHP and the ACA survey The latter was conducted according

deter-to the recommendations in Sawdeter-tooth Software’s recent ACA manual

Each respondent had to pass either the ACA or the AHP questionnaire to avoidlearning effects and to keep the time needed to complete the questionnaire withinacceptable limits Neither ACA nor AHP provide a general measure of predictivevalidity, which is usually quantified by presenting holdout tasks If the number ofattributes to be considered in a product evaluation problem is high, the use of hold-out stimuli is regularly accompanied by the risk of information overload (Herrmann

et al (2005)) The relevant set of attributes was determined for each respondent dividually to create a realistic choice setting Each respondent was shown reducedproduct stimuli consisting of his or her six most important attributes Accordingly,the predictive validity was measured by means of a computer administered holdouttask similar to the one proposed by Herrmann et al (2005)

in-Choice tasks including three holdout stimuli were presented to each respondentafter having completed the preference measurement task One of these alternativeswas the best option available for the respective respondent (based on an online es-timation of individual part-worth utilities carried out during the interview) The twoother stimuli were slight modifications of this best alternative Each one was gen-erated by randomly changing three attribute levels from the most preferred to thesecond or third most preferred level.

In the last part of the online questionnaire, each respondent was faced with his

or her individual profile of attribute importance estimates In this regard, the

corre-sponding question “Does the generated profile reflect your notion of attribute tance?" had to be answered on a 9-point rating scale ranging from “poor" (= 1) to

impor-“excellent" (= 9)

The respondents were invited to participate in the survey via a large public e-maildirectory For practical reasons we sent 50 % more invitations to the ACA than to theAHP survey We obtained 380 fully completed questionnaires for ACA and 204 forAHP In both cases, more than 40 % of those who entered the online interview alsocompleted it Chi-square homogeneity tests show that both samples are structurallyidentical with respect to socio-demographic variables

to both measures, namely ACR = 17 and R2= 77, the internal validity of our study

can be rated high To come up with a fair comparison, we accepted all completedquestionnaires and did not eliminate respondents from the samples on the basis of

ACA’s R2or AHP’s ACR.

As a first step in our empirical investigation, we compared the resulting erence structures on the aggregate level We transformed the part-worth utilities ofboth methods such that they sum up to zero for all levels of each attribute to facilitatedirect comparisons The attribute importances were transformed in both cases suchthat they sum up to one for each respondent Spearman’s rank correlation was used

pref-to contrast the convergent validity of AHP with ACA Table 2 provides the attributeimportances and the transformed part-worth utilities of both approaches The differ-ences regarding the part-worth utilities are rather small Although both methods areconceptually different, the obvious structural equality points to high convergent va-lidity The rank correlation between AHP and ACA part-worth utilities equals 90

In contrast, there are substantial differences between the attribute importances of

AHP and ACA on the aggregate level (r = −.08) To assess the factual quality of

at-tribute importances, we verified the present results by considering previous empirical

studies in the field of tourism In a recent study by Hamilton and Lau (2004) the cess to the sea or lake was ranked second among the 10 attributes considered in this study The importance of the corresponding attribute location in our study is higher

ac-for AHP than ac-for ACA which favors the values provided by the ac-former Analogously,

the attribute active sports (which corresponds to leisure activities in our study) was

rated as very important by only 6 % of the respondents in a survey by Study Group

“Vacation and Travelling" (FUR (2004)) On the other hand, the importance of the

attribute relaxation, which is similar to outside facilities in our study, was highly

ap-preciated Insofar, the AHP results are in line with the FUR study by awarding high

importance to outside facilities and lower importance to leisure activities.

To find an appropriate external criterion that allows to measure the validity ofthe resulting individual attribute importances is difficult We chose the respondents’individual perceptions as an indicator and measured the adequacy of the importance

Table 2 Average attribute importances and part-worth utilities

Attribute Importance Part-worths* Importance Part-worths*

-.03 (3) 06 (2) -.01 (3) 03 (2)Catering 12.17 -.19 (5) 03 (3) 13.29 -.07 (5) -.01 (3)

.12 (1) -.07 (4) 02 (2) -.04 (4)

Hotel facilities

Location 7.78 -.24 (2) 24 (1) 12.84 -.32 (2) 32 (1)Type of building 9.09 08 (2) -.22 (3) 8.36 -.03 (2) -.12 (3)

estimates in the last part of the questioning by means of a 9-point rating scale

ques-tion (see Secques-tion 3) Here, AHP was judged significantly better (p < 01) with an

average value of 7.3 compared to ACA with 6.68 This suggests that AHP yieldshigher congruence with the individual perceptions than ACA But since it is notclear to what extent respondents are really aware of their attribute importances, theexplanatory power of this indicator has not been fully established

The predictive accuracy of both methods was checked by comparing the overallutilities of the holdout stimuli with the actual choice in the presented holdout task

as explained in Section 3 Both methods were evaluated by two measures: The first choice hit rate equals the frequency with which a method correctly predicts the vaca-

tion package chosen by the respondents Here, AHP significantly outperforms ACA

with 83.33 % against 60.78 % (p < 01) The overall hit rate indicates how often a

method correctly predicts the rank order of the three holdout stimuli as stated by therespondents Taking into account that the respondents had to rank alternatives of theirevoked sets (i.e the best and two “near-best" alternatives) the predictive accuracy ofboth approaches is definitely satisfying Again, AHP significantly outperforms ACA

with an overall hit rate equal to 63.42 % compared to 43.94 % for the latter (p < 01) For comparison: random prediction would lead to an overall hit rate equal to 1¯6 All

in all, AHP shows a significantly higher predictive accuracy for products belonging

to the evoked set of the respondents than ACA

5 Conclusions and outlook

This paper presents an online implementation of AHP for consumer preference surement in the case of products with larger numbers of attributes As a first bench-mark, we empirically compared AHP with Sawtooth Software’s ACA in the domain

mea-of summer vacation packages While both methods yielded high values for internaland convergent validity, AHP significantly outperforms ACA regarding individuallytailored holdout tasks generated from the respondents’ evoked sets The results sug-gest AHP as a promising method for preference-driven new product development.Further empirical investigations are required to support the results presented here.These should include additional preference measurement approaches, such as SE orBridging CA (Green and Srinivasan (1990)) Moreover, the implication of differ-ent hierarchies have not been fully understood in AHP research (Pöyhönen et al.(2001)) While we conducted extensive pre-studies to come up with an expedienthierarchy, market researchers should be very carefully when structuring their deci-sion problems hierarchically The application of simple 3-level hierarchies focusing

on the main goal, attributes and levels only, and leaving out higher-level attributecategories might be beneficial These hierarchies would also be reasonable when theproduct evaluation problem cannot be broken down into ‘natural’ groups of attributecategories

FRICKER, S., GALESIC, M., TOURANGEAU, R and YAN, T (2005): An Experimental

Comparison of Web and Telephone Surveys, Public Opinion Quarterly, 69, 3, 370–392.

FUR (2004): Travel Analysis 2004 by Study Group “Vacation and Travelling", www.fur.de.GREEN, P.E and SRINIVASAN, V (1990): Conjoint Analysis in Marketing: New Develop-

ments with Implications for Research and Practice, Journal of Marketing, 54, 3–19.

HAMILTON, J.M and LAU, M.A (2004): The Role of Climate Information in Tourist nation Choice Decision-making, Working Paper FNU-56, Centre for Marine and ClimateResearch, Hamburg University

Desti-HAUSER, J.R and TOUBIA, O (2005): The Impact of Utility Balance and Endogenity in

Conjoint Analysis, Marketing Science, 24, 3, 498–507.

HERRMANN, A., SCHMIDT-GALLAS, D and HUBER, F (2005): Adaptive Conjont ysis: Understanding the Methodology and Assessing Reliability and Validity In: A

Anal-Gustafsson, A Herrmann and F Huber (Eds.): Conjoint Measurement: Methods and plications, Springer, Berlin, 253-278.

Ap-PÖYHÖNEN, M., VROLIJK, H and HÄMÄLÄINEN, R.P (2001): Behavioral and

Proce-dural Consequences of Structural Variation in Value Trees, European Journal of tional Research, 134, 1, 216–227.

Opera-SAATY, T.L (1980): The Analytic Hierarchy Process McGraw Hill, New York.

SAWTOOTH SOFTWARE (2003): SSI Web v3.5, Sawtooth Software

SAWTOOTH SOFTWARE (2005): Report on Conjoint Analysis Usage among Sawtooth ware Customers, www.sawtoothsoftware.com, accessed March 2007

Soft-SCHOLL, A., MANTHEY, L., HELM, R and STEINER, M (2005): Solving MultiattributeDesign Problems with Analytic Hierarchy Process and Conjoint Analysis: An Empirical

Comparison European Journal of Operational Research, 164, 760–777.

SCHOLZ, S.W and DECKER, R (2007): Measuring the Impact of Wood Species on sumer Preferences for Wooden Furniture by Means of the Analytic Hierarchy Process

Con-Forest Products Journal, 57, 3, 23–28.

SCHOLZ, S.W., MEISSNER, M and WAGNER, R (2006): Robust Preference Measurement:

A Simulation Study of Erroneous and Ambiguous Judgement’s Impact on AHP and

Con-joint Analysis In: H.-O Haasis, H Kopfer and J Schönberger (Eds.): OR Proceedings

2005, Springer, Berlin, 613–618.

Management of Customer Equity and How to Obtain the Most from Spatial Customer Equity Potentials

Klaus Thiel and Daniel ProbstDepartment of Analytical Customer Relationship Management

CRM, T-Online, T-Com, DTAG

T-Online Allee 1, 64295 Darmstadt, Germany

Abstract This paper will show, how the usage of an early-warning system, which has been

developed and implemented for a big internet service provider, can detect customer equity tentials respectively risks and how to use this information to launch special customer treatmentdepending on strategic customer control dimensions in order to increase customer equity Thestrategic customer control dimensions are: customer lifetime value, customer lifecycle andcustomer behaviour types The development of the customer control dimensions depends onthe availability of relevant customer data Thus, from the huge amount of available customerdata, relevant attributes have been selected In order to reduce complexity and use standard-ised processes the raw-data is aggregated, for example into clusters We will demonstrate bymeans of a real-life example the detection of spatial customer equity risks and the launch ofcustomer equity increasing treatment using the early-warning system in interaction with thementioned strategic customer control dimensions

po-1 Introduction1

In the b2c-sector a continuing increase in competition, growing customer tions, variety seeking and an erosion of margin can be observed The solution of suc-cessful enterprises is a paradigm change from product-centred to customer-centredorganisations combined with the long-run objective of customer equity (CE) max-imisation

expecta-In this paper we will show both theoretically and by means of a real-life example,the detection of CE risks at a very early stage as well as the launch of CE increasingtreatment using the early-warning system (EWS) in interaction with the strategic cus-tomer control dimensions (SCCD) The SCCD are customer lifetime value (CLV),customer lifecycle (CLC) and customer behaviour types (CBT)

1The content of this article originates from various projects All projects have been cuted under the leadership or substantial cooperation of the aCRM department, with thedepartment head Klaus Thiel