Personalization of the search process in tourism [Elektronische Ressource] / Sven Döring
140 Pages
English

Personalization of the search process in tourism [Elektronische Ressource] / Sven Döring

Downloading requires you to have access to the YouScribe library
Learn all about the services we offer

Description

PERSONAL IZAT ION OF THE SEARCH PROCESS IN TOUR ISM Doctoral Thesi s Dipl .-Inf. Sven DöringFaculty of Applied Co mputer ScienceUniversity of AugsburgDoering@In formatik.Uni-Augsburg .de© Copyr ight 2008. All r ights reserved. 2Examiners: Prof. Dr. Werner KießlingProf. Dr. Bernhard Möl lerDay of oral exam ination: 09.05 .20083Abstra ct The combination of trav el and touri sm represents the leading domain for application s in B2C e-comm erce. Thus, it deserves high est attent ion. Since m ost people only have a ver y limited nu mber of vaca tion days each y ear, they have l earned t o be more dem anding abou t their trips . More and m ore they ask for bette r-per sonalized travel products inst ead of standard packages design ed by tour ist operators. Due to insuffic ient search engines and the lack of personaliza-tion, however, arranging a trip on current online tr avel portals is of ten not as easy as it should be. Even f or rather s traightforward sc enarios, s earch ing and booking a s uit able travel package can be tedious and might often take longer than 1 hour. In order to prov ide good sal es experi-ences a nd custom-tailored produc ts similar to the ones co mpetent human travel agents c an of-fer, a personalized search approach for online travel portals has been ov erdue for some time.

Subjects

Informations

Published by
Published 01 January 2008
Reads 27
Language English
Document size 1 MB

PERSONAL IZAT ION OF
THE SEARCH PROCESS
IN TOUR ISM
Doctoral Thesi s

Dipl .-Inf. Sven Döring
Faculty of Applied Co mputer Science
University of Augsburg
Doering@In formatik.Uni-Augsburg .de
© Copyr ight 2008. All r ights reserved. 2
Examiners: Prof. Dr. Werner Kießling
Prof. Dr. Bernhard Möl ler
Day of oral exam ination: 09.05 .20083
Abstra ct
The combination of trav el and touri sm represents the leading domain for application s in B2C
e-comm erce. Thus, it deserves high est attent ion. Since m ost people only have a ver y limited
nu mber of vaca tion days each y ear, they have l earned t o be more dem anding abou t their trips .
More and m ore they ask for bette r-per sonalized travel products inst ead of standard packages
design ed by tour ist operators. Due to insuffic ient search engines and the lack of personaliza-
tion, however, arranging a trip on current online tr avel portals is of ten not as easy as it should
be. Even f or rather s traightforward sc enarios, s earch ing and booking a s uit able travel package
can be tedious and might often take longer than 1 hour. In order to prov ide good sal es experi-
ences a nd custom-tailored produc ts similar to the ones co mpetent human travel agents c an of-
fer, a personalized search approach for online travel portals has been ov erdue for some time.
This thesis , therefore, presents a novel personalized search process de livering travel produc ts
exactly tailored to custo mers with respec t to their situat ions and pr efer ences .
In a first step, a nove l m odel for the search process in elec tron ic com merce will be intro-
duced . A deep p ersonalization o f the s earch wi ll be p rovided b y dividing t he proc ess in to four
stages, namely Preference Analys is & Modeling, Search Interface, Query Processing, and
Presentation. The m ain part of this thesis , wi ll then appl y the new mo del to the tour ism do-
m ain, i.e . each s tep of the sea rch wil l be examined in the con text o f tou rism. A situat ion mod-
el ad equate ly adjusted to the tourism domain wil l then provide each stage of the search pro-
cess wi th addi tiona l situat ional knowl edge. Based on this , several essential co mponents for a
domain spec ifi c search in tou rism wil l be introduced according ly: a new prefe rence construc-
tor dealing with typ ical pri ce-qua lity trad eoffs, a smart pref erence elicitation process support -
ing custo mers who have to find an optim al depa rtur e airport, the co mposition and evalua tion
of da tabase que ries support ing the interpla y of ind ividual and global preferenc es, and an ap-
propriate adapta tion of search interfa ce and produc t presentation . Moreove r, by us ing prefer-
ence search technolog ies as underlyi ng basis for the search itself, best alternatives can be de-
livered in case there i s no perfec t match.
Severa l nove l softwar e components for a personali zed search proce ss in tourism have come
Tinto exis tence i n the con text o f this thes is, e.g ., t he person ali zed protot ype COSIMA . The in -
terpla y of these co mponents with ex isting preferenc e components wil l be examined and eval-
uated by means of nu merous use ca se scena rios at the end of this work. It wi ll be demonstrat -
ed t hat by a proper c o mbination of these components, custo m-tailored t ravel products with re-
spec t to prefer ences and situa tions can be found and presented to the custo mer in an intui tive ,
fast and m ore co mfortable manner than before. 5
Ack nowledgments
This work would not have been possible wi thout the valu able suppo rt of several people I am
particularly gratefu l to. At t he c hair of databases and i nformation system s of the University of
Augsburg (Germany) m y doctoral adviser , Prof. Dr. Werner Kießling, prov ided m e wi th an
inspiring vi ew into a world ful l of preferences. I would like to thank him for his support and
good advice during m y research .
I am deepl y gratefu l for the incentive and jo yful work wi th m y colle agues in the department.
In particular , I would like to thank Stefan Hol land and Ste fan Fisch er for a smooth start into
the Preference World, Timotheus Preisinger , Ma rkus Endres , and Alfons Huhn for the good
cooperation during va rious projects . I am also espec ial ly grateful for the suppor t of Prof. Dr.
Bernhard Möl ler and Anna Schwartz .
The work at FORSIP, the Bava rian Research Cooperation for Situated , Individualized and
Personalized Hu man-Com puter Interaction, offered me insights into va rious interesting re-
search f ields. I would l ike to t hank all the people a t FORSIP, I had th e pleasure to work wi th.
Specia l thanks are due to Al ex Mango ld f or a last-minute, linguistic rev iew of m y work.
Finally, I am ver y grateful to Ste fanie Leis tner for he r support and pa tienc e.7
Contents
1 Introduction.......................................................................................................................9
2 Trave l Search in Tourism ..............................................................................................13
2.1 Customer's Experience with a Typical Trav el Portal.................................................13
2.2 An Overv iew of Existing Online Travel Portals........................................................16
2.2.1 Exped ia................................................................................................................17
2.2.2 Tr avelocity..........................................................................................................18
2.2.3 TUI .....................................................................................................................18
2.2.4 Tr aveltainm ent and '5vorFlug'.............................................................................19
2.2.5 Summ ary ...........................................................................................................20
2.3 Objectives of this Thesis ...........................................................................................21
3 Foundations of Prefer ences Rev isited............................................................................25
3.1 Modeling Preferenc es ...............................................................................................25
3.1.1 Base Pre es.................................................................................................27
3.1.2 Complex Prefe rences30
3.1.3 SV-Semantics......................................................................................................32
3.2 The Preference Framework .......................................................................................34
3.2.1 Pr efer ence Se arch ................................................................................................35
3.2.2 Si tuat ed Pre ference Mod el and Pr efer ence Repos ito ry........................................36
3.2.3 Personalized Presentat ion of Quer y Results........................................................37
4 Tailoring a Persona lized Search f or Tour ism ...............................................................39
4.1 Personalizat ion of the Sea rch P rocess in E-Com merce ..............................................39
4.1.1 Comm o n Prefer ences in Elect ronic Co m merce ...................................................39
4.1.2 Des ign Princip les f or a Personalized Search Proces s...........................................42
4.1.3 Search M odel .......................................................................................................43
4.2 Situat ion M odel ing for the Sea rch P rocess in Tourism45
4.2.1 Si tuat ed Ent ity- Relationship -Model for Tour ism ................................................46
4.2.2 Tourism-Rela ted Pr efer ence Repos ito ry .............................................................48
4.3 Tradeoff Preferenc e Const ructo r...............................................................................49
4.3.1 Def inition o f the Prefe rence Construc tor55
4.3.2 Complexity and Performance Considera tions ......................................................61
4.3.3 Quality Valuat ion ...............................................................................................62
4.4 Smart Preferenc e El icitat ion......................................................................................64
4.4.1 Pr efer ence E licitation Based on Information Int egration.....................................66
4.4.2 Smart Preferenc e El icitat ion in Conclusion.........................................................75
4.5 Advanced Preference Quer y Process ing ...................................................................76
4.5 .1 Pr efer ence Query Expansion Approach ...............................................................79
4.5 .2 Algorithm and Co mplexit y .................................................................................92
4.6 The Concepts in Retrospect.......................................................................................948
5 Prefe rence Based Co mponents for Touri sm..................................................................95
5.1 Histo ry of COSIMA .................................................................................................95
5.2 Novel Com ponents for a Deepl y Personalized Search in Touri sm ............................97
5.2.1 Airport Finder and Tr avel Reco m mender............................................................97
5.2.2 Personalization of the Search Interface: Visua lization Co mponent .....................99
T5.2.3 Advanced Pr efer ence Query Proc essing : COSIMA .........................................100
5.3 Putt ing the Pieces Together: Int erpla y of Co mponents ............................................101
5.4 Typical Use Case Scen arios.....................................................................................103
6 Achievements and Related Work.................................................................................109
6.1 Search Process Model..............................................................................................109
6.2 Situ ation al I nfluences111
6.3 Int egration o f Heterogeneous Information Sources.................................................112
6.4 Search Process i n Tour ism ......................................................................................113
7 Su mmary and Out look..................................................................................................117
7.1 Su m mary of this Thesis ...........................................................................................117
7.2 Future Work ............................................................................................................118
Bibliog raphy....................................................................................................................121
Appendix A......................................................................................................................1351Introduc tion
The travel and tour ism industry makes up a conside rable branch of our economy. Accord ing
1
to the World Travel & Tourism Counci l , almost 11% of the world wide GDP is represen ted
by travel and tou rism. Besides , travel a nd tou rism is the l ead ing app lica tion field in B2C elec -
tron ic com m erce; it represen ts nearly 50% of the t ota l B2C turnover ( [Wer03]). The tr avel in-
dustry and its products have rather spec ific featur es, which ma kes them predes tined for online
dis tribu tion : the product is a confidence good and has a short-liv ing natur e, custo mer dec i-
sions are mo stly based on information, and the industry is high ly networked , based on the
world-wide coope ration of di ffe rent typ es of stakeholde rs ( [Wer03, Ben06]).
2 3
Low- cost airlines like Ryanair or EasyJ et have gained huge significance in rec ent years.
They strong ly rely on their onl ine portals in order to sell ticke ts. Nevertheles s, the re are still a
lot of people who prefe r to consult a hum an employee in a travel agenc y instead of using the
internet for book ing or organizing a journe y. Although the travel industry alread y stands for
an important part of elec troni c comm erce, there is still a lot of room for improvem ent. The
ability to conduc t a task such as search ing is absolut ely cen tral for travel websites ([Ben06]).
However, due to techn ica l problems, incomprehensible interfac es, and insuffic ient search en-
gines, ave rage custo mers are often ov er-s trained when arranging a va cation on the internet.
This might be one reason wh y there is still little evidence of elec troni c m arkets leading in-
var iabl y to lower search cos ts ([ÖK03]) .
Since mo st people have on ly a very limited nu mber of vaca tion da ys each year, they have be-
come more demanding abou t their trips. Mo re and mo re they ask for better personalized trav -
el p roducts i nst ead of s tandard pa ckages des igned b y tou rist operators . This might be one rea -
son for the emerging concep t of dynamic packaging. A vacation pack age combines hotels
with flights and/o r rental cars in one price . Cus to mers become their own trav el agent and
build their travel packages themselves. This has been introduced to the tour ism industry as an
approach to ach ieve a competitive advantage prov iding custo mers with flexible travel pack -
ages ([CL07]). The author s also state that produc t packages should be custo mized based on
the requirements spec ified by customers, who, in add ition, are no longer content wi th just
find ing the lowest price. Custom ers are looking to take contro l and identify the perfect trip
([Gro07]). But this ev en worsens the problems mentioned above, since existing online portals
1 www.wttc.travel
2 www.r ya nair. com
3 www.eas yj et.com10 Personalizat ion of the Sea rch P rocess in Tourism
4
like Expedia are sufficien t for simple problems like inquiring a fligh t fro m A to B with spe-
cific hard constraints. Mor e complex problems, on the other hand , require more intelligent
sys tem s ([Wer03]).
Besides tourists traveling dur ing their vaca tion , the re are bu siness travelers who also need hu -
man-friend lie r search engine s in ord er to reduce the ir valuable time for find ing and booking a
bus iness trip on the internet. Travel expenses make up som e of the largest expenses in com -
panies, second on ly to labor and IT. At the same time, they are an excell ent target for a cost
reduction which can be induced by optimizing individual processe s such as search ing and
book ing. The selec tion of the 'righ t' hotel, for exa mple, is insufficien tly desc ribed by price
alone, as choosing a suitable room for a bus iness trip is a non-t rivial task. A nu mber of a
hotel's cha ract eristics (e.g ., fac ilities) as we ll as its loc ation m ust be taken into consid eration
([NMT06]) . In add ition, there are also several cr iteria related to the travelers and thei r opti -
mal accom m odation with respec t to individual pref erence s, e.g ., the quality of the hotel. Note
that this comprises only on e part of the trav el package ; taking other parts such as fligh ts and
rental cars into account wi ll m ake the dec ision-making process even more complex.
Only rec ently, the trav el and tourism indust ry ha s learned to take advan tage of search engin es
as an important marketing argum ent in prom otion. There was a widespread advertising slo-
5
gan of Fly.de , a German online portal for fligh ts (see Figure 1.1): “Fliegst Du schon oder
su chst Du noch?” , which m eans something along the lin es of 'Are you flyi ng or still tryi ng (to
find a n appropriate flight)?'.
Figure 1.1. Fly.de advertising sloga n
The idea of bu ying travel package s vi a on line travel port als is simple. Cus to mers would like
to have at leas t the serv ice they would have when dire ctly contacting a hum an employer in a
travel agenc y. This m eans that custom ers want to be treated individual ly, accord ing to his or
her wishes and situ ation . Gener ic, 'on e-si ze-fits-al l' sea rch engines will not su ffi ce in tourism.
Therefore, good search engines for tour ists have to be adapted to this domain. Furthermore,
4 www.expedia.com
5 www.fl y.de

)