Download Using Self-Organizing Maps and K

C. H. Lien et al. / Asian Journal of Management and Humanity Sciences, Vol. 1, No. 1, pp. 1-15, 2006
Capturing and Evaluating Segments: Using
Self-Organizing Maps and K-Means in Market
Segmentation
CHE-HUI LIEN1,*, ALEX RAMIREZ2, AND GEORGE H. HAINES2
1
Department of Management, Thompson Rivers University, Canada
2
Eric Sprott School of Business, Carleton University, Canada
ABSTRACT
Ma
r
ke
ts
e
g
me
nt
a
t
i
o
ni
sav
i
t
a
lp
a
r
to
fa
no
r
g
a
n
i
z
a
t
i
o
n’
sma
r
ke
t
i
ngbe
c
a
u
s
ei
tprovides the
fundamental framework necessary for effective marketing efforts. In recent years, due to their high
performance in engineering, artificial neural networks have also been applied in management research.
Self-organizing maps, a technique of unsupervised neural networks, are often used for clustering or
dimensional reduction. This study employs a modified two-stage approach (SOMs and K-means) to
group customers, compares the performance between the tandem approach and direct K-means
clustering, and tests for the existence of clusters and segments. The test results show that a media
promotion variable would be a basis for segmentation. Based on the segmenting results, a marketing
c
o
mmu
ni
c
a
t
i
o
ns
t
r
a
t
e
g
yi
spr
e
s
e
nt
e
dt
oc
o
pewi
t
hc
us
t
o
me
r
s
’e
x
pe
c
t
a
t
i
o
ns.
Key words: market segmentation, cluster analysis, data mining, neural networks, self-organizing maps.
1. INTRODUCTION
Since the pioneering research of Wendell Smith (1956), the concept of market
segmentation has been one of the most pervasive in both the marketing academic
l
i
t
e
r
a
t
u
r
ea
n
dpr
a
c
t
i
c
e(
Ku
o,Ho,& Hu
,2002)
.I
nt
oda
y
’
sc
ompe
t
i
t
i
v
ema
r
k
e
t
pl
a
c
e
,
locating and effectively targeting unique market segments enables a company to
understand the wants and needs of its customers.
For several decades, statistical cluster analysis has been successfully used in
market segmentation (Green & Krieger, 1995). Recently, due to an increase in
computer power and a decrease in computer cost, a great deal of interest and effort
have been directed towards using neural networks (NNs) in business practice,
which were once reserved for multivariate statistical analysis. In marketing, the
major application of NNs is on market segmentation. Along with the evolution of
data mining techniques, Self-organizing maps (SOMs) have been used in
determining clusters and is an alternative approach to statistical clustering
techniques (Bigus, 1996; Fish, Barness, & Aiken, 1995; Kuo et al., 2002;
Venugopal & Baets, 1994). An SOM provides a mapping from a high-dimensional
input data into the lower dimensional output maps. A distinguished feature of the
SOM is that it preserves the topology of the input data from the high-dimensional
input space onto the output map in such a way that the relative distances between
input data are more or less preserved (Garson, 1998; Thanakorn, 2003).
Although a number of clustering methods have been presented to solve the
market segmentation problem, the importance of testing the validity of clusters is
* Corresponding author. E-mail: [email protected].
1
C. H. Lien et al. / Asian Journal of Management and Humanity Sciences, Vol. 1, No. 1, pp. 1-15, 2006
frequently ignored by marketing researchers (Arnold, 1979; Engelman & Hartigan,
1969; Pilling, Crosby, & Ellen, 1991; Punj & Steward, 1983). If identified clusters
are inadequately validated, such clustering results can be considered random, i.e.,
devoid of a meaningful structure (Pilling et al., 1991; Punj & Steward, 1983). The
neglect of the test for clusters could lead to the inclusion of clusters that really do
not exist.
Segments are defined as groups showing some patterns of similarity and that
differ significantly in their response to the relevant marketing variables (Sommers
& Barnes, 2003). Clusters identified that differ in some way, such as attitudes or
demographics, but not in behavior are not true segments (Massy & Frank, 1965).
Kuo et al. (2002) proposed a modified two-stage approach (SOMs and
K-means) and, based on simulated data, they found that it is slightly better (lower
rate of misclassification) than the traditional statistical clustering method. But Kuo
et al. (2002) used a tandem approach, based on a preliminary factor analysis
followed by a clustering of rotated, standardized factor scores, in the K-means
procedure. Green and Krieger (1995) criticized the tandem approach and they
argued that the performance of the tandem approach was not as good as direct
K-means clustering. In addition, Kuo et al. (2002) neither tested for the existence of
clusters nor evaluated whether their clusters were segments. More empirical studies
are needed to support the better performance of the novel two-stage approach.
This study uses practical data from the home heating system market and
a
ppl
i
e
s Ku
oe
ta
l
.
’
st
wo-stage method (SOMs and K-means) in market
segmentation, compares the performance of direct K-means clustering with the
tandem approach, and tests for clusters and segments.
The remainder of this paper di
s
c
u
s
s
e
s SOMs
’a
ppl
i
c
a
t
i
on
si
n ma
r
k
e
t
segmentation and the test method of clusters and segments, followed by a review of
the methodology used in this study. The section after that presents our results. The
final section gives a summary and a discussion about the findings of the study.
2. REVIEW OF RELATED LITERATURE
2.1 SOMs in Market Segmentation
The general idea of segmentation is to group items that are similar
(homogeneous) (Kuo et al., 2002; Smith, 1956). Traditionally, statistical clustering
techniques have been the common tools for market segmentation (Green & Krieger,
1995). Punj and Steward (1983) suggested that the integration of a hierarchical
a
ppr
oa
c
h
,s
u
c
ha
sWa
r
d’
smi
n
i
mum v
a
r
i
a
n
c
e
,a
l
on
gwi
t
han
on
-hierarchical one,
such as the K-means, can provide a better answer than using either a hierarchical or
a non-hierarchical method alone. Their approach is called the two-stage approach.
The term neural networks arose from artificial intelligence research, which
attempted to understand and model brain behavior (Berry & Linoff, 2000; Berson,
Smith, & Thearling, 2000). Self-organizing maps are feed-forward, unsupervised
neural networks and were developed by Kohonen (Bigus, 1996; Kohonen, 2001).
Feed-forward networks are used in situations where we can bring all of the
2
C. H. Lien et al. / Asian Journal of Management and Humanity Sciences, Vol. 1, No. 1, pp. 1-15, 2006
information to bear on a problem at once, and we can present it to the neural
networks (Bigus, 1996). Unsupervised learning means that clustering proceeds
without knowing (a priori) the number of clusters of the input data, and the network
organizes and learns the input data unsupervised (Bigus, 1996). The SOM consists
of two layers of processing units, an input layer fully connected to an output layer.
There is no hidden layer.
Vanugopal and Baets (1994) argued that unsupervised neural networks,
including SOMs and the adaptive resonance theory, could be used in determining
clusters. Their network for segmentation was conceptually developed. Fish et al.
(1995) also suggested that unsupervised neural networks could be utilized for
market segmentation. In the housing market, Kauko, Hoomineijer, & Hakfoort
(2002) examined the application of neural networks to housing market
segmentation in Helsinki, Finland. Their study shows how it is possible to identify
various dimensions of housing segments by uncovering patterns in the data set as
well as the classification ability of SOM-LVQ (learning vector quantization). Kuo
e
ta
l(
2002
)
,ba
s
e
donPu
n
ja
n
dSt
e
wa
r
d’
s(
1983)r
e
s
e
a
r
c
h
,pr
opos
e
damodi
f
i
e
d
two-stage approach, which first used self-organizing maps to determine the number
of clusters and then employed the K-means method to find the final solution. Kuo
et al. (2002) used simulated data and found that their proposed two-stage approach
outperformed the conventional two-stage method. The main reason was that the
first stage of the conventional two-stage clustering method always involved the
hierarchical methods. One of their shortcomings was the aspect of non-recovery:
once an observation has been assigned to a cluster, it should not be moved at all
(Kuo et al., 2002). However, SOMs are a kind of learning algorithm, which can
c
on
t
i
nu
a
l
l
yu
pda
t
e
,orr
e
a
s
s
i
gnt
h
eobs
e
r
v
a
t
i
ont
ot
h
ec
l
os
e
s
tc
l
us
t
e
r
.I
nKu
oe
ta
l
.
’
s
study, they employed the tandem approach in the K-means procedure and the
tandem approach was criticized in that it could distort the original cluster structure
(Green & Krieger, 1995). It is necessary to examine the non-tandem approach, such
as direct K-means clustering, and to evaluate its performance.
2.2 Test of Clusters and Segments
An appropriate test for clusters appears to be one, which takes into account
the objective of cluster analysis, i.e., to minimize the within-group variance and
maximize the between-group variance (Arnold, 1979). Among the test methods, a
method suggested by Arnold (1979) appears to be better than other methods
(
Pi
l
l
i
n
ge
ta
l
.
,1991;Pu
n
j&St
e
wa
r
d,1983)
.I
nAr
n
ol
d’
st
e
s
t
,t
h
es
i
gn
i
f
i
c
a
n
c
eoft
h
e
cluster solution is tested by comparing the calculated value of the C test statistics3
to the values which would be expected if the data were drawn from either a
unimodal or a uniform distribution, i.e., no basis for clusters (Arnold, 1979).
There is some consensus in marketing research on what makes a segmentation
solution a good one (Bacon, 2002; Wedel & Kamakura, 2000). The important
criteria are: identifiability, accessibility, substantiality, stability, responsiveness,
3
C = log (max (|T| / |W|)) where T represents the total scatter matrix, and W the pooled within-group matrix.
3
C. H. Lien et al. / Asian Journal of Management and Humanity Sciences, Vol. 1, No. 1, pp. 1-15, 2006
and actionability (Bacon, 2002; Wedel & Kamakura, 2000). These criteria are in
terms of usefulness for managerial decision-making and are not easy to evaluate
(Bacon, 2002). They are also independent of the technology used to obtain them
(Bacon, 2002). Therefore, they provide a useful set of goals for segmentation
solutions developed using neural network techniques (Bacon, 2002). Few studies
attempted to assess segmentation solutions with these criteria (Wedel & Kamakura,
2000). In this paper, we employ one criterion –responsiveness (consumers in
different segments should behave differently toward marketing programs directed
at them) –to evaluate whether clusters are segments. The major reason for
choosing responsiveness as the criterion is that the data we used are cross-sectional.
Other criteria, e.g., stability, need longitudinal data. Responsiveness is consistent
wi
t
hMa
s
s
ya
n
dFr
a
nk’
s(
1965)a
r
g
ument, i.e., different segments should have
different promotional or marketing variables elasticities.
3. METHODOLOGY
3.1 Research Sample
This study used previously collected data (Ratchford & Haines, 1986) for
clustering analysis. Data were collected from a US sample of Market Facts
Consumer Mail Panel in January-February 1983. Because weather is freezing in
winter, home heating systems are important in some states (e.g., New York State)
in the US. In the 1980s, basically there are five types of home heating system: gas,
oil, electric, heat pump and solar. The families in the US can choose one of them to
pr
odu
c
eh
e
a
ti
nt
h
eh
ous
e
.I
n Ra
t
c
h
f
or
d & Ha
i
n
e
s
’
sr
e
s
e
a
r
c
h
,t
h
e
yi
de
n
t
i
f
i
e
d
fourteen perceptual attributes as the criteria for rating the five types of home
heating systems. They also run factor analysis among the fourteen attributes and
found that the resulting three factors (cleanliness, reliability, and cost) explaining
approximately 60% of the variation between attributes. However, they did not
perform segmentation among the consumers.
We focus on consumers planning on buying a house within the next five years.
This research defines the respondents without planning to replace home heating
systems in the next five years, but planning on buying a house within the next five
years as the potential market. When buying a new house, one type of home heating
system must be considered in the purchase. The subjects included both males and
females, above the age of 18. 433 responses were obtained for the potential market.
3.2 Research Hypotheses
3.2.1 Test of clusters
According to the procedure for testing for clusters described by Arnold (1979),
there are two sets in the procedure. The first tests the null hypothesis that the
population entities tend to concentrate at one point (i.e., one normally distributed)
against the alternative hypothesis that its entities are either uniformly distributed or
grouped into clusters. If the null hypothesis is rejected at some level of confidence,
4
C. H. Lien et al. / Asian Journal of Management and Humanity Sciences, Vol. 1, No. 1, pp. 1-15, 2006
a second test is made. The second tests the null hypothesis that the population
entities are uniformly distributed against the alternative hypothesis that they form
two or more clusters. The two null hypotheses are:
H1: the population entities tend to concentrate at one point;
H2: the population entities are uniformly distributed.
Rejection of H1 and H2 indicates that there are clusters in the population.
3.2.2 Test of segments
The null hypothesis is:
H3: the marketing variables elasticities among clusters are the same.
Rejection of H3 means the clusters are segments.
3.3 Questionnaire
Young, Ott, and Feigin (1978) argued that segmentation based on benefits
desired is usually the most meaningful type to use from a marketing standpoint as it
directly facilitates product planning, positioning, and advertising communication.
The main strength of benefit segmentation is that the benefits sought by customers
will lead to a causal relationship to future purchase behaviour (Minhas & Jacobs,
1996). This study uses benefit variables to group customers. In Ratchford and
Ha
i
n
e
s
’
sr
e
s
e
a
r
c
h(1986), the questionnaire solicited information on five types of
home heating systems and each system type was rated for each of the fourteen
attributes (benefit variables, see Table 1).
Table 1. Fourteen benefit variables
Fourteen Benefit Variables
Reliability against breakdown
Future availability of fuel supply
Floor space required for the system
Efficiency of converting fuel source into heat
Cleanliness of operating the system
Ease of conversion to another system
Absence of fumes and odors
Absence of pollution
Availability of professional servicing
Noise-free operation of the system
Safety of the system
Initial purchase price
Warranty protection
Annual operating cost
Purchase intentions were measured by the fol
l
owi
ngqu
e
s
t
i
on
:“
I
fy
ouwe
r
e
buy
i
ngah
e
a
t
i
n
gs
y
s
t
e
mf
ory
ou
rh
ome
,pl
e
a
s
e‘
x’t
h
eon
et
y
pet
h
a
ty
ouwou
l
dbe
mos
tl
i
k
e
l
yt
opu
r
c
h
a
s
e
.
”Re
s
pon
s
e
swe
r
el
i
mi
t
e
dt
ot
h
ef
i
v
es
y
s
t
e
mt
y
pe
s
.Ot
h
e
r
questions involve the knowledge of heating system, rating all the methods of
central heating, marking the preference for one method of home heating over the
other, media, demographic information, etc. The measurement scale in the
questionnaire is 7-point scale, ranged from not very important (=1) to very
important (=7).
5
C. H. Lien et al. / Asian Journal of Management and Humanity Sciences, Vol. 1, No. 1, pp. 1-15, 2006
3.4 Generating Clusters: Using SOMs
The Kohonen algorithm can be summarized in the following steps (Bigus,
1996; Garson, 1998; Kohonen, 2001) and the self-organizing maps parameters used
in our study are listed in Table 2.
(1) Neuron weights are initialized to random values.
(2) Data representation. When clustering data with neural networks, it is
standard practice to scale (or normalize) the input data to a range of zero
to one (Bigus, 1996; Garson, 1998). This study uses the sigmoid function4
to transfer input data to a range of zero to one.
(3) Setting the learning rate. The learning rate is used to control the step size
in the adjustment of the connection weights. This study sets the learning
rate as 0.1. The learning rate will decrease over time.
(4) Setting the neighborhood. We set a 10-by-10 output layer. The
neighborhood is 10 in this study.
(5) Presetting the number of training epochs (training cycle). An epoch is the
number of training data sets presented to the model between updates of
neural weights (Garson, 1998). The purpose of training is that neural
networks have the ability to learn and recognize any important pattern or
relationship for a set of data (Nguyen & Ramirez, 1998). So far, there is no
rule for determining the number of learning epochs. The only way is trial
and error (Kuo et al., 2002). In our research, the training epochs for
potential market are 80. At the 80th epoch, the network converges to an
average total minimum distance.
(6) Data points are input to the net, selected at random.
(7) Determine which neuron is the least distant from the presented
observations. This is the neuron whose weight vector is closest to the input
vector from the current observations, measured in Euclidean distance.
(8) Weights of neurons in the neighborhood of the winning neuron are
adjusted in value to become closer to the value of the winner. The
neighbourhood starts out widely defined but decreases spatially as learning
iterations proceed, eventually reaching zero (that is, only the weights of
the winning neuron are adjusted). When the neighbourhood drops to one,
the convergence phase begins.
In reviewing the SOM literature (Berry & Linoff, 2000; Bigus, 1996; Kuo et
al., 2002), usually 70% of the population used for training, and 30% of the
population used for testing were suggested. In the 433 observations of potential
market, a sample size of 300 is used for training and the remaining 133
observations are used for testing (see Table 2).
4
The sigmoid function is defined as a strictly increasing function that exhibits smoothness and asymptotic properties.
One popular sigmoid function is: f(x) = 1/[1+ exp (-x)].
6
C. H. Lien et al. / Asian Journal of Management and Humanity Sciences, Vol. 1, No. 1, pp. 1-15, 2006
Table 2. The self-organizing maps parameters
The self-organizing maps parameters
[ 0] Number of Input Unit (<20): 14
[ 1] Number of Rows (=Columns) of Output Unit (about 4-10, <20): 10
[ 2] Number of Train Examples: 300
[ 3] Number of Test Examples: 133
[ 4] Number of Train Cycles (about 20-200): 80
[ 5] Number of Test Period (about 1-10): 1
[ 6] Using Batch Learn (Yes=1,No=0, usually 0): 0
[ 7] Using Learned Weights (Yes=1,No=0, usually 0): 0
[ 8] Range of Weights (0.1-0.5, usually 0.3): 3.000e-01
[ 9] Random Seed (0.1-0.9, usually 0.456): 4.560e-01
[10] Learn Rate (0.01-0.3, usually 0.1): 1.000e-01
[11] Learn Rate Reduced Factor (0.9-1.0, usually 0.95): 9.500e-01
[12] Learn Rate Minimum Bound (0.01-0.1, usually 0.01): 1.000e-02
[13] Radius (about = Number of Rows): 1.000e+01
[14] Radius Reduced Factor (0.9-1.0, usually 0.95): 9.500e-01
[15] Radius Minimum Bound (about 0.1): 1.000e-01
This study uses three different software packages – PCNeuron, SAS
Enterprise, and Neuroshell to implement clustering. When we set the same
parameters, such as the learning rate = 0.1, to implement SOMs in the three
software packages, the visualization clustering results are consistent among the
three software packages –three clusters in the potential market. Although the
aggregate results –in terms of identifying the number of possible clusters –are the
same, the same individual may not always be assigned to the same cluster. After
running SOMs, the number of clusters is used to implement the K-means algorithm.
4. RESULTS
4.1 Self-Organizing Maps
Through the implementation of SOMs by the PCNeuron, the SOMs clustering
result is consistent between training and testing samples, showing that there are
three clusters in the potential market (see Figure 1).
4.2 K-means Analysis
From the previous section, the number of clusters (= 3) is used to implement
direct K-means clustering. As mentioned in the questionnaire, the original 14
benefit variables are directly used as input variables and the clustering results show
that there are 143 members in cluster 1, 273 members in cluster 2, and 17 members
in cluster 3. The members in cluster 1 concern more in the reliability and safety of
the home heating system (means (centroids): reliability against breakdown = 6.13,
safety of the system = 5.92, warranty protection = 5.16, floor space required for the
7
C. H. Lien et al. / Asian Journal of Management and Humanity Sciences, Vol. 1, No. 1, pp. 1-15, 2006
system = 4.48, availability of professional servicing = 4.64, future availability of
fuel supply = 4.2, ease of conversion to another system = 4.43), the members in
cluster 2 put much attention to price (initial purchase price= 6.76, annual operating
cost = 6.59, efficiency of converting fuel source into heat = 6.5), and the members
in cluster 3 care about the cleanliness and quietness of the home heating system
(cleanliness of operating the system = 6.74, absence of pollution =6.43, absence of
fumes and odors = 6.83, noise-free operation of the system = 5.57).
A Wi
l
k’
sl
a
mbdav
a
l
u
ei
sus
e
dt
oc
ompa
r
et
h
epe
r
f
or
ma
n
c
eofc
l
u
s
t
e
r
i
n
g
.
Wi
l
k
’
sLa
mbdav
a
l
u
ei
st
h
er
a
t
i
ooft
h
ewi
t
h
i
n-group variance (SSwithin ) to the total
variance (SStotal ) (Stevens, 2002). The total variance is the sum of the
between-group variance and the within-g
r
ou
pv
a
r
i
a
n
c
e
.A Wi
l
k
’
sl
a
mbdav
a
l
u
e
closest to zero implies that the source of total variance is from the between-group
variance instead of from the within-group v
a
r
i
a
n
c
e
.Th
es
ma
l
l
e
rt
h
eWi
l
k’
sl
a
mbda
v
a
l
u
e
,t
h
ebe
t
t
e
rt
h
ec
l
u
s
t
e
r
i
ngr
e
s
u
l
t
.Th
ec
a
l
c
u
l
a
t
e
dWi
l
k’
sl
a
mbdav
a
l
u
e
sa
r
e
:
Wi
l
k
’
sλ
=SSwithin / SStotal = 19.31/ 1691.182 = 0.0114 (direct K-means);
Wi
l
k
’
sλ
=SSwithin / SStotal = 2.904/ 131.378 =0.0221 (tandem approach).
Because 0.0114 < 0.0221, the performance of the direct K-means clustering is
better than the performance of the tandem approach.
4-3 Test of Clusters
The best fitting functions for the unimodal and uniform distribution (Arnold,
1979) are as follows:
Cunimodal = e0.06239 g0.95242 p0.21011 α0.13389 / N0.29723 = 0.6379;
Cuniform = e0.08664 g0.89510 p0.13896 α0.16356 / N0.23672 = 0.6859,
where g = the number of clusters = 3; p = the number of attributes = 14; N = the
n
umbe
rofe
n
t
i
t
i
e
s=433;α=the level of significance = 0.1. Since the calculated
value of test statistics C = log (1691.182/19.31) = 1.942, which exceeds the critical
value for Cunimodal and Cuniform, both null hypotheses (H1 and H2) are rejected.
Therefore, this study can conclude that the clusters exist in the potential market.
4-4 Test of Segments
Because the study wants to estimate marketing variables (price, media
promotion, and total knowledge of home heating systems) elasticities, all variables
(dependent and independent) will be transformed to logarithms before estimation so
that the regression function can be written as a classical linear regression. Since the
variables are logarithms, it is possible to interpret the regression coefficients as
elasticities. The Chow test (Doran, 1989) can be used to test the equality of
regression coefficients across two or more sets. If the test is significant, the null
hypothesis (H3) will be rejected, and we can conclude that the clusters are
segments.
8
C. H. Lien et al. / Asian Journal of Management and Humanity Sciences, Vol. 1, No. 1, pp. 1-15, 2006
30
Series1
Series2
25
Series3
20
Series4
Series5
15
Series6
10
Series7
5
S9
S5
0
1
3
S1
5
7
Series8
Series9
Series10
9
(a)
Series1
15
Series2
Series3
10
Series4
Series5
Series6
5
Series7
S7
0
S1
1
4
7
Series8
Series9
Series10
10
(b)
Figure 1. Visualization of clusters in the potential market. (a) Visualization of clusters for
training samples in the potential market. (b) Visualization of clusters for testing
samples in the potential market.
9
C. H. Lien et al. / Asian Journal of Management and Humanity Sciences, Vol. 1, No. 1, pp. 1-15, 2006
The study uses purchase intention as dependent variable and price, media
promotion, total knowledge of the home heating systems as independent variables.
In price variables, there are ten variables: initial purchase price of oil, electric, solar,
gas, heat pump and annual operating costs of oil, electric, solar, gas, heat pump. As
per media promotion variable, it is defined as where customers see or hear anything
about heat pumps as a method of home heating systems from media (radio,
television, newspaper, magazine, and flyer or handbill) in the past 6 months.
Customers can choose all tools that apply to see or hear heat pumps. As per total
knowledge of the home heating systems, it stands for the total score of the
knowledge of oil, electric, solar, gas, and heat pump systems. After deleting
observations with missing information regarding home heating system questions in
the potential market, there are 130 valid members in cluster 1, 248 valid members
in cluster 2, and 11 valid members in cluster 3. Through the stepwise or forward
selection procedures, 6 independent variables (total knowledge of the home heating
systems, initial purchase price of heat pump, annual operating costs of electric,
annual operating costs of solar, annual operating costs of gas, media promotion) are
identified as good predictors. The backward selection procedure identifies 7
predictors. On the grounds of parsimony (Stevens, 2002), we might prefer the 6
predictors selected by the stepwise or forward procedure, especially because the
adjusted R2 for the stepwise and backward procedure are quite close (0.535 and
0.539). This model also yields a significant F-test of overall model fit (see Table 3).
Table 3. Results of multiple-regression in the potential market
Dependent Variable: Purchase Intention
Independent Variables
Intercept
Total knowledge of the home heating systems
Initial purchase price of heat pump
Annual operating costs of electric
Annual operating costs of solar
Annual operating costs of gas
Media promotion
F-test
Adjusted R2
Coefficient
1.324
0.952
-.174
.209
.184
-.408
.248
17.819*
0.535
Sig.
.000
.000
.040
.002
.004
.000
.000
Note. * P< 0.05
The ANOVA results of multiple-regression in the potential market are shown
in Table 4. From Table 4, the F-ratio (Chow test) is:
F=
[37.474 (12.615 23.431 0.173] /(6 1)
= 1.845
(12.615 23.431 0.173) /[130 248 11 2(6 1)]
F = 1.845 < F(7, 375, 0.95) =3.
2
4
5(
s
e
tα=0.
05)
.
10
C. H. Lien et al. / Asian Journal of Management and Humanity Sciences, Vol. 1, No. 1, pp. 1-15, 2006
The F statistics shows that the test is not significant and we can conclude that
the clusters are not segments.
Table 4. ANOVA of multiple-regression
Source
Cluster 1
Regression
Residual
Total
Cluster 2
Regression
Residual
Total
Cluster 3
Regression
Residual
Total
Entire Data Set
Regression
Residual
Total
Sum of Squares
DF
Mean Squares
F-test
3.829
12.615
16.444
6
123
129
0.638
0.103
6.2*
6.895
23.431
30.325
6
241
247
1.149
0.097
11.84*
0.829
0.173
1.001
6
4
10
0.138
0.043
3.2
10.523
37.474
47.997
6
382
388
1.754
0.098
17.9*
Note. * P< 0.05
For the potential market, only 11 valid members in cluster 3 and the sample
size is very small. This raises a natural question about the power of the test.
Therefore, the study further compares cluster 1 and cluster 2, and finds the Chow
test is still not significant (F = 0.949 < F(7, 364, 0.95) = 3.25). Because the Chow test
tests all the variables together, the effects of the variables that do have the same
c
oe
f
f
i
c
i
e
n
t
sc
ou
l
dpos
s
i
bl
y“
s
wa
mp”t
h
ee
f
f
e
c
t
sofonl
yon
eort
wooft
h
ev
a
r
i
a
bl
e
s
with the different coefficients. If such differences did exist, variables found with
different coefficients would be the basis for segmentation. The media promotion
variable is found with a significant t-test result (t = 2.09 > t (376, 0.95) =1.645) (the
regression coefficients are different between cluster 1 and cluster 2), and it would
be a basis for segmentation.
According to the definition of media promotion variable described in this
section, this study calculates the frequencies of radio, television, newspaper,
magazine, and flyer or handbill chosen by customers (Table 5) and we find that in
cluster 1, most customers see or hear about heat pumps as a method of the home
heating system from radio (35.3%), newspaper (45.4%), and magazine (51.5%). In
cluster 2, most customers see or hear about heat pumps from newspaper (41.1%)
and magazine (50.4%).
11
C. H. Lien et al. / Asian Journal of Management and Humanity Sciences, Vol. 1, No. 1, pp. 1-15, 2006
Table 5. Frequencies of radio, TV, newspaper, magazine, and flyer in Cluster 1 and Cluster
2
Cluster 1
N
%
Cluster 2
N
%
Radio
TV
Newspaper
Magazine
Flyer or Handbill
46
35.3
24
18.5
59
45.4
67
51.5
16
12.3
19
7.6
51
20.6
102
41.1
125
50.4
35
14.1
5. CONCLUSIONS
This study employs, for the first time, the direct K-means procedure in the
modified two-stage clustering approach (SOMs and K-means). The result reveals
that direct K-means clustering outperforms the tandem approach. It could be
complementary to Kuo et al. (2002) modified two-stage clustering (they employ the
tandem approach in the K-means clustering) and provides a better solution to the
problem of market segmentation. In addition, the use of neural networks is not just
procedure specific (in terms of results), but program specific. Three different
software packages are utilized in our research and the results have been discussed.
There does not seem to be much use at the test that ensures groups are really
c
l
u
s
t
e
r
s
.Ou
rr
e
s
e
a
r
c
he
mpl
oy
sAr
n
ol
d’
sa
ppr
oa
c
ht
ot
e
s
tt
h
ec
l
us
t
e
r
sa
n
da
v
oi
d
falling into the trap that we always find clusters. The test at segmentation (the
Ch
ow t
e
s
t
)s
h
owst
h
a
t“
bl
i
n
dr
e
l
i
a
n
c
e
”ont
h
eCh
ow t
e
s
ti
sn
otag
oodi
de
awh
e
n
seeking to see if segments really exist.
Advertising in newspapers and magazines as well as promotion through radio
programs would be effective to communicate with customers of segment 1. In
segment 2, advertising or promoting mainly in newspapers and magazines could be
effective marketing communication with customers of segment 2. A major research
limitation is that due to commercial secrecy, our study project did not get any help
from Canadian companies offering historical data. The authors use previously
collected data from Ratchford and Haines (1986). Because the data were collected
in 1983, if the results were to be put to immediate practical use by the home heating
system company, it would have to be assumed that customer benefit perception has
remained stable over the intervening years.
REFERENCES
Arnold, S. J. (1979). A Test for Clusters. Journal of Marketing Research, 16(4):
545-551.
Bacon, L. D. (2002). Handbook of Data Mining and Knowledge Discovery. Oxford,
UK: Oxford University Press.
Berry, M. J., & Linoff, G. (2000). Mastering data mining: The art and science of
customer relationship management. New York, USA: John Wiley & Sons.
12
C. H. Lien et al. / Asian Journal of Management and Humanity Sciences, Vol. 1, No. 1, pp. 1-15, 2006
Berson, A., Smith, S., & Thearling, K. (2000). Building Data Mining Applications
for CRM. New York, USA: McGraw-Hill.
Bigus, J. (1996). Data Mining with Neural Networks. New York, USA:
McGraw-Hill.
Doran, H. (1989). Applied Regression Analysis in Econometrics. New York, USA:
Marcel Dekker, Inc.
Engelman, L., & Hartigan, J. (1969). Percentage Points of a Test for Clusters.
Journal of the American Statistical Association, 64, 1647-1648.
Fish, K., Barnes, J., & Aiken, M. (1995). Artificial neural networks: A new
methodology for industrialmarket segmentation, Industrial Marketing
Management, 24, 431-438.
Garson, G. (1998). Neural networks: An introduction guide for social scientists.
Thousand Oaks, California, USA: SAGE Publications.
Green, P. E., & Krieger, A. M. (1995). Alternative Approaches to Cluster-Based
Market Segmentation. Journal of the Market Research Society, 37(3), 221-239.
Kauko, T., Hoomineijer, P., & Hakfoort, J. (2002). Capturing housing market
segmentation: An alternative approach based on neural networks modeling.
Housing Studies, 17(6), 875-894.
Kohonen, T. (2001). Self-oorganizing maps (3rd ed.). Springer Series in
Information Sciences. Berlin, Germany: Springer-Verlag.
Kuo, R. J., Ho, L. M., & Hu, C. M. (2002). Cluster analysis in industrial
segmentation through artificial neural networks. Computers and Industrial
Engineering, 42, 391-399.
Massy, W., & Frank, R. (1965). Short Term Price and Dealing Effects in Selected
Market Segments. Journal of Marketing Research, 2(2), 171-185.
Minhas R., & Jacobs E. (1996). Benefit segmentation by factor analysis: An
improved method of targeting customers for financial services. International
Journal of Bank Marketing, 14(3), 3-13.
Nguyen D., & Ramirez A. (1998). The emerging position of artificial networks as a
prime intelligent technology for strategic decision support system. Information
Systems, 19(4), 14-22.
Pilling, B., Crosby, L., & Ellen, P. (1991). Using benefit segmentation to influence
environmental legislation: A bottle bill application. Journal of Public Policy
and Marketing, 10(2), 28-46.
Punj, G., & Stewart, D. (1983). Cluster analysis in marketing research: Review and
suggestions for application. Journal of Marketing Research, 20(2), 134-148.
Ratchford, B., & Haines, G. (1986). A Study of consumer behavior in a product
class which contains new technologies. Contemporary Research in Marketing,
1, 403-424.
Smith, W. (1956). Product differentiation and market segmentation as alternative
marketing strategies. Journal of Marketing, 21(3), 3-8.
Sommers, M., & Barnes, J. (2003). Introduction to marketing (10th ed.). New
York, USA: McGraw-Hill.
Stevens, J. (2002). Applied multivariate statistics for the social science. Mahwah,
New Jersey, USA: Lawrence Erlbaum Associates.
13
C. H. Lien et al. / Asian Journal of Management and Humanity Sciences, Vol. 1, No. 1, pp. 1-15, 2006
Thanakorn N. (2003). Data mining applications for self-organizing maps.
Unpublished Ph.D. dissertation. Rensselaer Polytechnic Institute.
Venugopal, V., & Baets, W. (1994). Neural networks and their applications in
marketing management. journal of systems management, 45(9), 16-21.
Wedel, M., & Kamakura, W. (2000). Market segmentation: conceptual and
methodological foundations (2nd ed.). Norwell, Massachusetts, USA: Kluwer
Academic Publishers.
Young, S., Ott, L., & Feigin, B. (1978). Some practical consideration in market
segmentation. Journal of Marketing Research, 15 (3), 405-413.
Che-hui Lien received a B.B.A from National
Cheng Kung University, Tainan, Taiwan and an MBA
degree on international trade from National Cheng Chi
University, Taipei, Taiwan. He holds a Ph.D. in
management (marketing) from Carleton University,
Ottawa, Canada.
He is currently an Assistant Professor in Marketing
at Thompson Rivers University in Kamloops, BC,
Canada. His research interests are database marketing,
services marketing, data mining, relationship marketing,
CRM, international marketing and marketing research.
Alex Ramirez received a B.Sc. with Honours
from ITESM (Instituto Tecnologico y de Estudios
Superiores de Monterrey) Monterrey Campus in his native
Mexico and a master's degree on industrial engineering
and operations research from Syracuse University in New
York, U. S. A. He holds a Ph.D. in administration
(information systems) from Concordia University,
Montreal, Canada.
He is currently an Assistant Professor in
information systems at the Eric Sprott School of Business
in Ottawa, Ontario, Canada. He was a Visiting Researcher
at the Westminster School of Business, in London, UK during the academic year
2005-2006. Currently he chairs the Academic Computing Committee and the
Curriculum Review Committee. His research interests are information systems
foundations; and evaluation and adoption of emerging technologies in
organizations: knowledge management systems, business intelligence, data
warehousing, data mining, and electronic commerce.
14
C. H. Lien et al. / Asian Journal of Management and Humanity Sciences, Vol. 1, No. 1, pp. 1-15, 2006
George H. Haines graduated from Massachusetts
Institute of Technology with an SB, Carnegie Institute of
Technology with an M.S. and Ph.D. He is currently a
Distinguished Research Professor at Carleton University.
His current research interests include the study of how the
financial marketplace works for small and medium sized
enterprises, with a focus on Canadian problems, and the
design of management games for education and training.
15