J. Serv. Sci. & Management, 2008,1: 244-250
Published Online December 2008 in SciRes (www.SciRP.org/journal/jssm)
245
Market Segmentation for Mobile TV Content on Public
Transportation by Integrating Innovation Adoption
Model and Lifestyle Theory
Chi-Chung Tao
Tamkang University, Taipei County, Taiwan, R.O.C.
Received June 20
th
, 2008; revised August 10
th
, 2008; accepted September 16
th
, 2008.
ABSTRACT
An integrated approach based on innovation adoption model and lifestyle theory for customer segmentation of mobile
TV content on public transportation using multivariate statistical analysis is proposed. Due to high daily trips and dif-
ferent train types Taiwan Railway Administration is chosen as the case study. Firstly, the content of mobile TV on the
train are identified as the segmentation variable and key factor facets for mobile TV content are renamed by using fac-
tor analysis. Then, the cluster analysis is used to classify customer groups which are named by analysis of variance
(ANOVA) and market segmentations are described with demographic, lifestyle and train patronage variables by using
cross analysis and Chi-squared independence tests. Finally, this paper discusses empirical results to provide valuable
implications for better mobile TV content marketing strategies in the future.
Keywords: mobile TV, market segmentation, multivariate statistical analysis
1. Introduction
Mobile TV has been widely discussed among different
players in the telecommunications and media industry.
Mobile operators, which face the saturation of voice ser-
vices and a declining ARPU (Average Revenue Per User),
hope that the TV concept in the mobile phone will be the
next killer application. According to analysts’ predictions,
mobile TV will become a service with significant market
size. The valuations of the global mobile TV market vary
across different analysts, from US$ 5.5 billion in 2009 to
US$ 28 billion in 2010. A list of different analysts’ pre-
dictions is shown in Table 1[1,2].
There are many trial projects confirming future success
of mobile TV worldwide using different technologies and
business models [3,4,5]. Currently three standards are
competing with each other: Digital Video Broadcasting
Transmission System for Handheld Terminals (DVB-H),
MediaFLO and Digital Multimedia Broadcasting (DMB)
Table 1. Global mobile TV market values
Analyst firm Year Prediction of market size
Datamonitor 2009 US$ 5.5 billion
Strategy Analytics 2009 US$ 6.4 billion
ABI Research 2010 US$ 27 billion
Pyramid Research 2010 US$ 13 billion – 28 billion
Frost & Sullivan 2011 US$ 8.1 billion
which are listed in Table 2[6]. AT&T has announced that
its MediaFLO based mobile TV service will be going live
in May 2008. While Qualcom is in a position to leverage
other technologies or use it for open access technologies
such as WiMAX or use them for mobile services. It looks
like that despite the EU having embraced a single stan-
dard for mobile TV, the US market will remain frag-
mented with multiple technologies.
According to mobile TV usage patterns identified in
many worldwide trials TV content must adjust to mobile
context of use [3,4,5]. Results of a Finnish study show
that mobile TV users spent approximately 20 minutes a
day watching mobile TV and more active users watched
between 30 to 40 minutes per session [7]. Typical usage
environments include transportation terminals (airport,
train station, bus stop, etc.), in the moving vehicles, work-
ing places or at home. It is also found that smaller screens
and the duration of usage may have significant influences
on the types of mobile TV content as well as the way us-
ers’ willingness to pay for mobile TV.
In Taiwan, a handheld TV experimental project was
launched in October 2006. There were five teams partici-
pating in mobile TV trials in North and South Taiwan. The
MediaFLO was tested in North Taiwan, while the DVB-H
was chosen for South Taiwan. The experiment is expected

Market Segmentation for Mobile TV Content on Public Transportation by Integrating 245
Innovation Adoption Model and Lifestyle Theory
Table 2. Overview of mobile TV solutions
Technology
Major market Standard Group lndustrial players Comparisons
DVB-H Europe, Asia, North
America, Australia DVB OMA
Nokia, BenQ - Siemens,
Motorola, Samsung, LG,
Alcatel, … etc.
IP based; 90% power saving design, either use
independent MUX, OR
T-DMB South Korea DAB LG, Samsung some proprietary; bit rate 1.5Mbps, non-IP based;
external antenna; less power saving; be con-
structed based on existing DAB network
S-DMB South Korea, Japan DAB LG, Samsung, Alcatel Some proprietary; non-IP based, big antenna;
power saving; satellite transmission with terres-
trial repeaters
MediaFLO North America Qualcomm Qualcomm & CDMA
manufacturers Proprietary; non-IP based; bandwidth efficiency
ISDB-T Japan only Proprietary
MBMS 3G service regions 3GPP, 3GPP2 Most 3G manufacturer Bit rate 345kbps~2Mbps, IP based; based on 3G
network; standardization still in progress
IPTV over
WiMAX
WiMAX service
regions WiMAX Forum
Intel…etc. standardization still in progress
to end no later than June 2008 and the formal licenses for
mobile TV operators will be permitted after NCC’s (Na-
tional Communications Commission) official evaluations.
Preliminary results of this experiment show that end-users
will use mobile TV to fill in gaps in their daily schedules:
waiting for the bus or subway, sitting in the train etc. In
these situations mobile TV competes with other possibili-
ties such as reading a book, listening to radio, playing a
mobile game or just watching out the window. In these
scenarios mobile TV might be an appealing choice, but
only if it does not inflict significant costs. On the other
hand, if the pricing is low enough, there might be quite
large audiences awaiting the launch of mobile TV services.
A logical choice might be to keep subscription prices as
low as possible, thus maximize the popularity of mobile
TV, and subsidize the lower subscription income with
higher advertising revenue. In summary, these five teams
together with three main telecommunication operators
(CHT, TMT, FET) reach a consensus that a user-centered
content design will contribute to future success of Tai-
wan’s mobile TV market.
To survive in competitive mobile TV services markets,
the operators need to determine who the target customers
are, what motivates them and why they pay for the mobile
TV content. This process is called market segmentation,
by which companies are able to understand their loyal
customers and concentrate their limited resources into
them. Although there are many studies covering critical
variables for market segmentation of mobile TV content,
until now only few papers focus on public transportation
systems with high speed, especially on the train [8]. Cur-
rently Taiwan Railway Administration’s (TRA) is con-
ducting a BOT project to implement mobile commerce
services on the train. In addition, TRA’s network is across
north and south Taiwan, content diversity of mobile TV
can be tested on the train countrywide. Theses advantages
may attract the five teams to focus on certain killer appli-
cations for mobile TV.
The concept of segmentation in mobile TV marketing
recognizes that consumers differ not only in the price they
will pay, but also in a wide range of benefits they expect
from the content. Good mobile TV content with compel-
ling value-added services are provided by tight business
and strategic partnership arrangements and by involving a
large number of companies, with each influencing other
parties in the value chain. Wang [9] identifies that power-
ful actors, such as carriers and the media industry’s con-
tent providers must agree on business models that support
the new ecosystem of mobile TV. Carlsson and Walden [7]
conclude that mobile TV content is the key factor to de-
termine the adoption and usage of mobile TV, especially
when traveling with public transportation to and from
work in order to relax or to keep up to date with the latest
news.
This paper is aimed on proposing an integrated ap-
proach for market segmentation of mobile TV content on
the train by integrating innovation adoption model and
lifestyle theory. Figure 1 shows the conceptual framework
of this integrated approach.
First, the mobile TV content is identified as the seg-
mentation variable. And key factor facets for mobile TV
content are redefined by using factor analysis. Then, the
cluster analysis is used to classify consumer groups which
are named by analysis of variance (ANOVA) and market
segmentations are described with demographic, lifestyle
and train patronage variables by using cross analysis and
Chi-squared independence tests. Finally, empirical results
are analyzed and the conclusion follows.
2. Literature Review
It is the first step for mobile TV companies to identify the

246 Chi-Chung Tao
Figure 1. Conceptual framework of the integrated approach
primary variables of market segmentation to understand
their customers’ requirements, attitudes and habits. These
primary variables called segmentation variables which can
be derived from the technological, demographical and
psychological, behavioral perspective, such as age, in-
come, gender, occupation, attributes of product or service,
personal interest, consumer awareness, perception. A re-
view of prior studies suggests the theoretical foundations
of the hypotheses formulations [10]. To achieve this goal,
this paper examines two prevalent theories (lifestyle the-
ory and innovation diffusion theory) for identifying indi-
vidual acceptance of mobile TV content on the train.
Lifestyle is a key factor in determining the adoption rate
for mobile TV content. Individuals will adopt given be-
havior patterns representative of their lifestyles, and as a
consequence will purchase different types of mobile TV
content. Many studies verified that behavioral variations
in purchases even if there is no question of a mix of socio-
demographical variables coming into play, lead to a need
for research into lifestyle as a potentially influential factor.
Market segmentation according to features of lifestyle
divides the market into segments based on activities, in-
terests, and opinions [11]. The lifestyle segmentation in
this paper is defined by including variables like activities,
referring to the way in which individuals spend their time
and money; interests, which are those things in their im-
mediate surroundings they consider more or less impor-
tant; and opinions, the view they have of themselves and
of the world around them. One of well-known lifestyle
analysis methods is proposed by Wind and Green [12].
First, factor-groups influencing lifestyle patterns are re-
duced by using factor analysis. Then, the cluster analysis
is used to classify consumer groups. The interrelationships
among lifestyle patterns and other consumer behavioral
variables are verified by cross analysis.
The innovation diffusion theory (IDT) is a well-known
theory proposed by Rogers [13]. The adoption of mobile
TV content on the train could be studied from the perspec-
tive of information technology innovations. IDT states
that diffusion of an innovation depends on five general
attributes including relative advantage, compatibility,
complexity, observability, and trial ability. These charac-
teristics are used to explain the user adoption and decision
making process. They are also used to predict the imple-
mentation of new technological innovations and clarify
how these variables interact with one another. The central
concept of innovation diffusion is ‘‘the process in which
an innovation is communicated through certain channels,
over time, among the members of a social system.’’ How-
ever, research has suggested that only the relative advan-
tage, compatibility, and complexity are consistently re-
lated to innovation adoption [14]. Relative advantage is
similar to perceived usefulness, whereas complexity is
similar to perceived ease of use. Compatibility is the de-
gree to which the innovation is perceived to be consistent
with the potential users’ existing values, previous experi-
ences, and needs [15]. High compatibility will lead to
preferable adoption. Holak and Lehmann modified
Rogers’s model for measuring relative advantages [16].
They provided empirical evidence that relative advantage
and compatibility “directly” affect consumers' purchase
intention. The two factors have received favoritism in
recent published articles in the area of product innovation.
The two innovation characteristics are also peculiar in that
they are defined in relation to existing products, while
others (complexity, trial ability, and observability) indi-
cate the innate characteristics of an innovation.
Originating in IDT research, different adopter groups
perceive innovations and thus behave differently. Miller
[17] finds that prior knowledge of potential adopters can
focus the use of resources to prevent an innovation from
failing. Innovation theory can be applied to identify the
attitudes and behavior of early adopters, as a dynamic
basis for a market segmentation model. Therefore,
Holak’s new product adoption model is used to examine
the acceptance of mobile TV for public transportation
users. Through the review of relevant literature, following
variables of three dimensions for mobile TV content seg-
mentation analysis shown in Table 3 are examined in this
paper [18,19,20]:
Table 3. Segmentation dimensions
Demographic Psychographic Behavioral
Age, occupation ,
gender, income,
religion, nationality,
education, marital
status, ethnicity
Values, lifestyle,
interests, opin-
ions, activities
Product use pat-
terns, perceived
benefits (value-
added, ease of use,
usability, useful-
ness)
Innovators
Early
Adopters
Early
Majority
Late
Majority
Laggards
User Groups
’
different Values on New
Product or Service
Demographical
Variables
Lifestyle
Variables
Consumer
Characteri
s
tics
Public
Transit
Patronage
Variables
Awar
eness of
Mobile TV
Content
Adoption of
Mobile TV
Content with
Different User
Groups

Market Segmentation for Mobile TV Content on Public Transportation by Integrating 247
Innovation Adoption Model and Lifestyle Theory
To identify links among what the mobile TV operators
would know about their customers and the bundles of con-
tent they could offer, clustering algorithms are generally
used as the primary methodology for market segmentation.
Clustering analysis techniques have been discussed in
details in the literature [21,22,23]. The most popular is k-
means algorithm which together with its modifications
was broadly reviewed by different authors [24,25,26]. It is
also found that algorithms using computational intelli-
gence did not show better results than k-means, the com-
binations of several algorithms are very often recom-
mended as the conclusion. Zakrwska and Murlewski [27]
investigated the shortcomings and advantages of three
algorithms of clustering analysis: k-means, two-step clus-
tering and density based spatial clustering of applications
with noise. Their numerical tests showed that k-means is
very efficient for large multidimensional data sets, how-
ever depends strongly on the choice of input parameter k.
However, it is not recommended in the case of data sets
with noise.
3. Research Methodology
3.1 Research Design
The framework of research design is shown in Figure 2.
After conducting an interview survey (face-to-face) with
questionnaires, the mobile TV content on the train is iden-
tified as the segmentation variable and key factor facets
for mobile TV content are redefined by using factor
analysis. Then, the cluster analysis (k-means) is used to
classify consumer groups which are named by ANOVA.
The market segmentations are described with content us-
age, demographic, lifestyle and train patronage variables
by using cross analysis and Chi-squared independence
tests. Finally, each segment market can be targeted with
precise customer characteristics and those results are used
as the starting point for providing the market strategies.
The first round survey with 90 questionnaires was
conducted on the three train types from Taipei station to
Taichung station from 22
nd
to 28
th
December 2007. After
reviewing preliminary results, some question items were
modified and new question items of lifestyle variables
were supplemented. The second round survey with 500
questionnaires was conducted on the train from 18
th
to
25
th
January 2008. Two types of handheld devices
NOKIA N77 and N92 were used to demonstrate mobile
TV content. The valid sample consisted of 462 respon-
dents.
3.2 Research Model and Hypotheses
The research model tested in this paper is shown in Figure
3. With this integrated approach the mobile TV content
may be regarded as segmentation variables including will-
ingness to use, time of usage, price of willingness to pay,
type of payment, incentives to take train. The demograph-
ical variables (gender, occupation, age, income, educa-
tion), lifestyle variables (knowledge-oriented, recreation-
oriented, high living quality-oriented, favorite informa-
tion-oriented, price sensitive-oriented, fashion-oriented),
train patronage variables (frequency, travel time, train
type, trip purpose) are chosen as descriptive variables to
depict customer characteristics.
The following hypotheses of the proposed constructs are
based on prior studies in the relevant literature [10,14,15]:
H
1a
: No significant difference exists between each seg-
ment and gender.
H
1b
: No significant difference exists between each seg-
ment and occupation.
H
1c
: No significant difference exists between each seg-
ment and age.
Figure 2. Framework of research design
Figure 3. Research model and hypotheses

248 Chi-Chung Tao
H
1d
: No significant difference exists between each seg-
ment and income.
H
1e
: No significant difference exists between each seg-
ment and education.
H
2a
: No significant difference exists between each seg-
ment and frequency to take train.
H
2b
: No significant difference exists between each seg-
ment and travel time on the train.
H
2c
: No significant difference exists between each seg-
ment and train type.
H
2d
: No significant difference exists between each seg-
ment and trip purpose.
H
3a
: No significant difference exists between each seg-
ment and willingness to pay for mobile TV content.
H
3b
: No significant difference exists between each seg-
ment and payment for mobile TV content by time or ac-
cess frequency.
H
3c
: No significant difference exists between each seg-
ment and price of willingness to pay.
H
3d
: No significant difference exists between each seg-
ment and willingness to take train much more due to mo-
bile TV content.
H
4a
: No significant difference exists between each seg-
ment and knowledge-oriented lifestyle.
H
4b
: No significant difference exists between each seg-
ment and recreation-oriented lifestyle.
H
4c
: No significant difference exists between each seg-
ment and high living quality-oriented lifestyle.
H
4d
: No significant difference exists between each seg-
ment and favorite information-oriented style.
H
4e
: No significant difference exists between each seg-
ment and price sensitive-oriented lifestyle.
H
4f
: No significant difference exists between each seg-
ment and fashion-oriented lifestyle.
4. Results Analysis
The descriptive statistics of demographical variables and
train patronage variables are summarized as follows:
1) A total of 48% of the respondents were male. The
age of most respondents was from 19 to 33 years old with
a total of 64%. 70% of them were college educated, 29%
were students and 55% working in service industry and
trade. 31% of respondents earn under US＄300 per month
and 21% earning US＄800 – 1000 per month.
2) 23% of total respondents take train at least one time
daily and 31% taking at least one time per week. 36% of
them were commuters for work and school, and 25% were
leisure or tourism travelers. The percentage of travel time
from 31-60 minutes was 33% due to short-trip distance. A
total of 34% and 36% of the respondents most often take
the EMU commuter train type and “Zhi-Chiang” express
train type, respectively.
Factor analysis utilized the principal axis factoring
method in order to identify underlying constructs. The
KMO measure of sampling adequacy and Barlett’s tests of
sphericity provided support for the validity of the factor
analysis of the data set. The varimax rotation facilitated
interpretability.
The major mobile TV content consists of seven factor
facets which together explained 72.1% of the total varia-
tion shown in Table 4.
Factor facets are thus named: films, news, documenta-
ries, sports, leisure & tourism, entertainment and drama.
Measure validation was also examined for internal consis-
tency by computing Cronbach’s α coefficient. The average
Cronbach’s α value was found to be greater than 0.8, in
accordance with Nunnally’s standard [28]. The same pro-
cedure of factor analysis for lifestyle variables was also
conducted to obtain following factor facets with a satisfied
average Cronbach’s α value 0.7: knowledge-oriented type,
recreation-oriented type, high living quality-oriented type,
favorite information-oriented type, price sensitive-
oriented type and fashion-oriented type.
Table 4. Results of mobile TV content by factor analysis
Initial Eigen-values Extraction Sums of Squared Loading Rotation Sums of Squared Loading
Component Total
% of Variance Cumulative % Total % of Variance Cumulative % Total % of Variance Cumulative %
1(Films) 12.71 27.80 27.80 12.71 27.80 9.32 23.21 23.21 23.21
2(News) 7.24 13.41 41.21 7.24 13.41 7.57 12.16 12.16 35.37
3(Documeutaries) 4.18 10.06 51.27 4.18 10.06 3.98 9.89 9.89 45.26
4(Sports) 2.26 7.02 58.29 2.26 7.02 3.70 9.34 9.34 54.60
5(Leisure & Tourism) 1.60 6.22 64.51 1.60 6.22 2.53 8.22 8.22 62.82
6(Entertainment) 1.25 4.55 69.05 1.25 4.55 1.95 6.19 6.19 69.02
7(Drama) 1.04 3.01 72.07 1.04 3.01 1.22 3.06 3.06 72.07
8(Weather forecasts) 0.99 2.79 74.85
9(Stock reporting) 0.97 2.59 77.44
10(Religion tallking) 0.936 2.516 79.96

Market Segmentation for Mobile TV Content on Public Transportation by Integrating 249
Innovation Adoption Model and Lifestyle Theory
Then based on cluster analysis of six factor scores for
lifestyle variables, the whole sample was classified into
three groups: Group 1 (Group of being fashion and opti-
mistic), Group 2 (Group of being leisured and living in-
formation-oriented) and Group 3 (Group of being conser-
vative and traditional). ANOVA was conducted and the
results reveal that there are significant differences among
three groups on each factor scores.
The hypothesized relationships in Figure 3 are tested
using ANOVA, cross analysis and Chi-squared test to
examine if there is enough evidence to infer that two re-
search variables are related. Results of hypotheses tests
with P-value, Chi-squared values or F-value are summarized in
Table 5.
As shown in Table 5, approximately 60% of all hy-
potheses are proven to have significant differences at the
95% level.
To identify characteristics and the structure of the mar-
ket segments, a summarized analysis was conducted to
compare results of mobile TV content, demographical,
train patronage and lifestyle variables including gender,
occupation, age, income, education, frequency to take
train, travel time on the train, train type, trip purpose, will-
ingness to pay for mobile TV content, payment by time or
frequency, price of willingness to pay, type of lifestyle,
favorite content and willingness to take train due to mo-
bile TV content. The summarized profiles are shown in
Table 6.
As to the key information in the target group, the gen-
der is male, monthly income is under US＄300, the age
is from 19 to 33 years old. They are high school and col-
lege students and spend 31-90 minutes on the train per trip.
They will pay either US＄0.25/ h or US＄0.5 for one us-
age during the travel time between 61 and 121 minutes
and over 121 minutes. Their lifestyle types include enter-
tainment-oriented, fashion-oriented, price sensitivity-
oriented. Their favorite content includes films, news, en-
tertainment, documentaries and leisure & tourism. They
are willing to take train much more if TRA offers more
interesting mobile TV content.
Table 5. Results of hypotheses tests
Hypothesis P-value
Chi-squared (x
2
)
/ F-value Accept
Reject
H
1a
0.005 10.502 (x
2
) √
H
1b
0.010 29.287 (x
2
) ×
H
1c
0.021 28.165 (x
2
) ×
H
1d
0.012 22.760 (x
2
) ×
H
1e
0.467 7.666 (x
2
) √
H
2a
0.596 4.602 (x
2
) √
H
2b
0.002 24.890 (x
2
) ×
H
2c
0.668 4.068 (x
2
) √
H
2d
0.043 11.946 (x
2
) √
H
3a
0.022 11.463 (x
2
) √
H
3b
0.096 4.692 (x
2
) √
H
3c
0.032 19.678 (x
2
) ×
H
3d
0.000 22.980 (x
2
) ×
H
4a
0.597 0.516 (F) √
H
4b
0.000 21.63 (F) ×
H
4c
0.020 11.37 (F) √
H
4d
0.179 1.72 (F) √
H
4e
0.048 2.05 (F) √
H
4f
0.000 5.76 (F) ×
Table 6. Profiles of market segments
Group 1 (fashion and optimistic) Group2 (leisured and living
information-oriented) Group3 (conservative and traditional)
Sample size 193 94 175
Grender Male＞Female Female＞Male Male=Female
Occupation College students Service industry and trade Uniformly distributed
Age 19-33 years old 14-42 years old Uniformly distributed
Average income per month Under US $ 300 US $ 800-100 US $ 800-1500
Education College and high school College High school
Frequency to take train No distinct segments No distinct segments No distinct segments
Travel time on the train More than 90 minutes 31-90 minutes 31-60 minutes
Train type Zhi-Chiang and EMU Zhi-Chiang and EMU Zhi-Chiang and EMU
Trip purpose Commuter, Homebound traveler,
leisure traveler Commuter, leisure traveler Unifornly distributed
Willingness to use mobile TV
Yes for travel time between 61 and
121 minutes and over 121 minutes
Yes for travel time between
61 and 121 miuutes and over
121 minutes
Yes for travel time over 121
minutes
Payment by time or frequency both Prefer by frequency Prefer by frequency
Price of willingness to pay US $ 0.25/h and US $ 0.5 for one usage US $ 0.5 for one usage US $ 0.3 for one usage
Type of lifestyle Entertainment-oriented, high living
quality-oriented and fashion-oriented
High living quality-oriented Price sensitive-oriented
Favorite content News, films, entertainment, documen-
taries News, films, leisure & tour- ism, entertainment No preference
Duration of watching mobile
TV content More than 30 minutes 20-30 minutes No preference
Willingness to take train due
to mobile TV content Yes for the travel time over 90 minutes Yes for the travel time over 90 minutes Yes for cheaper fares only

250 Chi-Chung Tao
5. Conclusions
This paper identifies the new primary factors for mobile
TV content on the train which may not found in the pre-
vious studies. Additionally, this paper proposes a concise
framework of research methodology for market segmen-
tation and user preferences for mobile TV content on the
train.
However, the proposed approach is applicable to the
case of “on the train”. The case of “railway networks”
which fully represents mobile TV content anywhere and
anytime in stations and trains needs to be researched fur-
ther. It is also recognized that many other facets of indi-
vidual differences (e.g., psychological type, cognitive
processing skills etc.) may be candidate variables for
lifestyle consideration.
REFERENCES
[1] N. Holland, et al., “Rescuing 3G with Mobile TV: Buins-
ess Models and Monetizing 3G,” Pyramid Research,
March 2006.
[2] Wireless World Forum, “Mobile Youth 06 Video,”
mobileYouth06–part one, July 2006.
[3] J. Trefzger, “Mobile TV-launch in germany-challenges
and implications,” working paper No. 209, Institute for
Broadcasting Economics, Cologne University, Germany,
2005.
[4] S. Orgad, “How will mobile tv transform viewer’s experi-
ence and change advertising,” Final report, Dept. of Me-
dia and Communications, London School of Economics
and Political Science, November 2006.
[5] QuickPlayMedia, “Mobile TV and video survey 2008,”
Toronto, Canada, 2008.
[6] M. P. Shih, “Analysis of mobile TV and its key success
factors: from the perspective of mobile operator,” Pro-
ceedings of International Symposium on HDTV and Mo-
bile TV, Taipei, Taiwan, 2007.
[7] C. Carlsson and P. Walden, “ Mobile TV-to live or die by
content,” Proceedings of the 40th Hawaii International
Conference on System Sciences, IEEE, Hawaii, USA,
2007.
[8] C. M. Tan and C. C. Wong, “Mobile broadband race:
Friend or foe,” Proceedings of the International Confer-
ence on Mobile Business, IEEE, ICMB’06, 2006.
[9] G. Wang, “ Mobile TV value chain and operator strate-
gies,” Master’s Thesis, Dept. of Communication Systems,
School of Information and Communication Technology,
KTH, Finland, February 2007.
[10] T. M. Lee and J. K. Jun, “Contextual perceived usefulness?
toward an understanding of mobile commerce acceptance”,
Proceedings of the International Conference on Mobile
Business, IEEE, ICMB’05, 2005.
[11] J. T. Plummer, “The concept and application of lifestyle
segmentation,” Journal of Marketing, pp. 33-74, January
1974.
[12] Y. Wind and P. E. Green, “Some conceptual measurement
and analytical problem in life style research,” Life style
and Psychographics, Chicago, AMA, 1974.
[13] E. Rogers, Diffusion of Innovation, Free Press, New York,
1962.
[14] J. H. Wu and S. C. Wang, “What drives mobile commerce?
An empirical evaluation of the revised technology accep-
tance model,” Information & Management 42, pp.719–
729, 2005.
[15] Y. Shin, H. Jeon, and M. Choi, “Analysis of the consumer
preferences toward m-commerce applications based on an
empirical study,” Proceedings of International Conference
on Hybrid Information Technology, IEEE, ICHIT’06,
2006.
[16] S. L. Holak and D. R. Lehmann, “Purchase intentions and
the dimensions of innovation: An exploratory model,”
Journal of Product Innovation Management, 7 (1), pp. 59-
73, 1990.
[17] R. N. Miller, “Target marketing,” Multinational Market-
ing , Vol. 13, No. 10, 1993.
[18] A. L. Gilbert and J. D. Kendall, “A marketing model for
mobile wireless services,” Proceedings of the 36th Hawaii
International Conference on System Science (HICSS’03),
2003.
[19] H. H. Lin and Y. S. Wang, “Predicting consumer inten-
tion to use mobile commerce in taiwan,” Proceedings of
the International Conference on Mobile Business
(ICMB’05), 2005.
[20] H. Feng, T. Hoegler, and W. Stucky, “Exploring the criti-
cal success factors for mobile commerce,” Proceedings of
the International Conference on Mobile Business
(ICMB’06), 2006.
[21] C. C. Aggarwal, C. Procopiuc, J. S. Wolf, P. S. Yu, and J.
S. Park, “Fast algorithms for projected clustering,” Pro-
ceedings of SIGMOD Conference, Philadelphia, 1999.
[22] A. K. Jain, M. N. Murty, and P. J. Flynn, “Data clustering:
A review,” ACM Computing Surveys, Vol. 31, No. 3,
September 1999.
[23] M. Zait and H. Messatfa, “A comparative study of cluster-
ing methods,” FGCS Journal, Special Issue on Data Min-
ing, 1997.
[24] P. V. Balakrishnan, M. C. Cooper, V. S. Jacob, and P. A.
Lewis, “Comparative performance of the fscl neural net
and k-means algorithm for market segmentation,” Euro-
pean Journal of Operational Research, No. 93, 1996.
[25] H. Hruschka and M. Natter, “Comparing performance of
feedforward neural nets and k-means for cluster-based
market segmentation,” European Journal of Operational
Research, No. 114, 1999.
[26] C. Y. Tsai and C. C. Chiu, “A purchased-based market
segmentation methodology”, Expert Systems with Appli-
cations, No. 27, 2004.
[27] D. Zakrzewska and J. Murlewski, “Clustering algorithms
for bank customer segmentation”, Proceedings of the 5th
International Conference on Intelligent Systems Design
and Applications (ISDA’05), 2005.
[28] J. C. Nunnally, Psychometric Theory, McGraw-Hill, New
York, 1967.