Transcript 投影片 1

Data mining for shopping centres customer knowledge-management
framework
Dennis, C., Marsland, D., Cockett, T.(2001)
Journal of Knowledge Management. . Vol. 5, Iss. 4; pp. 368-374
授課教師: 許素華博士
學生
: S92660005黃永智
S92660014呂曉康
S92660017李峻賢
日期
: 2004/03/29
Agenda






Introduction
Exploratory Study
Results
Models of Relative Spend
Discussion and Conclusion
About K-Means
1
Introduction(1/2)

Knowledge is a fundamental factor behind an
enterprise's success





Management using knowledge-based computer systems
and networks
Management of intellectual (human) capital
All knowledge activities affecting success
Richards et al. (1998) argue that success is founded
on "a continuous dialogue with users, leading to a real
understanding".
For retailers the key ... is to establish data warehouses
to improve and manage customer relationships
(Teresko, 1999)
2
Introduction(2/2)




Incorporating data mining and customer
database aspects within a framework of
knowledge management can help increase
knowledge value.
Sharing information
Loyalty schemes
The objective of retail data mining schemes
has been to identify subgroups

Shopping and Service motivations
3
Exploratory Study



The results are from a survey of 287
respondents at six shopping centres
Determine which specific attributes of
shopping centres were most associated with
spend for subgroups of shoppers
Convenience sample –(weekdays, 10.30am to 3.30pm )


Unstructured Interviews
Least squares regression
4
Results
Conventional demographics(1/6)

Females vs males

The significant attributes for females were grouped around
two factors:


Shopping: "selection of merchandise"
Experience: "friendly atmosphere"
6
Conventional demographics(2/6)

Upper vs Lower socio-economic groups


ABC1 (managerial, administrative, professional,
supervisory and clerical)
C2DE (manual workers and pensioners)
7
Conventional demographics(3/6)

Higher vs Lower income groups
8
Conventional demographics(4/6)

Older vs Younger shoppers
9
Conventional demographics(5/6)

Shoppers travelling by car vs Public transport
10
Conventional demographics(6/6)

Service importance vs Shops importance
11
Cluster analysis

Shoppers motivated by
the "importance" of
"shops" vs "service“

A cluster analysis (SPSS
K-means) based on
"importance" scores has
identified distinct
subgroups sharing
particular needs or wants.
12
Compare Service & Shops Groups

“Service" shoppers were in a slightly higher
characters then “shops” shoppers gropus
Socio-economic group (63 % ABC1s vs. 59 %)
 Income (60 % ￡20,000 per year + vs. 53 %)
 Age (42 percent 45 + vs. 33 percent)
 Traveled by car (90 percent vs. 52 percent)

13
Models of Relative Spend
for "shops":
11 Spend = 19.4 + 0.70 X Attractiveness -0.21 X Distance.
14
Relationship Between Attractiveness & Sales
Turnover
15
Discussion and Conclusion

Target high-spending "service" shoppers.


Local loyalty cards are applicable and cost-effective
for cities and regional shopping centres



Increase their spend by 10%, equivalent to over a 3% rise
in total sales.
Car park membership scheme for in-town centres
Knowledge management network between retailers
and the centre would be a further stage
Most successful shopping centres are those where
“Active marketing" and “Proactive management" are
a feature
16
About K-Means
17
The K-Means Algorithm
1.
2.
3.
4.
5.
Choose a value for K, the total number of clusters.
Randomly choose K points as cluster centers.
Assign the remaining instances to their closest
cluster center.
Calculate a new cluster center for each cluster.
Repeat steps 3-5 until the cluster centers do not
change.
An Example Using K-Means
Table 3.6
• K-Means Input Values
Instance
X
Y
1
2
3
4
5
6
1.0
1.0
2.0
2.0
3.0
5.0
1.5
4.5
1.5
3.5
2.5
6.0
f(x)
7
6
5
4
3
2
1
0
x
0
1
2
3
4
Figure 3.6 A coordinate mapping of the
data in Table 3.6
5
6
Table 3.7 • Several Applications of the K-Means Algorithm (K = 2)
Outcome
Cluster Centers
Cluster Points
1
(2.67,4.67)
2, 4, 6
Squared Error
14.50
2
(2.00,1.83)
1, 3, 5
(1.5,1.5)
1, 3
(2.75,4.125)
2, 4, 5, 6
(1.8,2.7)
1, 2, 3, 4, 5
15.94
3
9.60
(5,6)
6
f(x)
7
6
5
4
3
2
1
0
x
0
1
2
3
4
Figure 3.7 A K-Means clustering of the
data in Table 3.6 (K = 2)
5
6
General Considerations

Requires real-valued data.
 We must select the number of clusters present in
the data.
 Works best when the clusters in the data are
of approximately equal size.
 Attribute significance cannot be determined.
 Lacks explanation capabilities.
簡報完畢 敬請指導
25