Download Study of Stock Customer Relationship Management Model Based on

Study of Stock Customer Relationship Management Model Based on
Data Mining
HUANG Feixue1, LI Zhijie2
1 Department of Economics, Dalian University of Technology, Dalian , P.R..China, 116024
2 Department of Computer Science and Engineering, Dalian University of Technology, Dalian , P.R..China, 116024
(E-mail:[email protected])
Abstract This study’s objective is to solve mining knowledge issue from different sources of distributed,
formatted or unformatted data with stock customer relationship management(CRM). The issue of is
formulated as data mining based on WEB. In particular, this paper presented a CRM model for data
mining of WEB based on software architecture of a novel infrastructure and the components of model are
discussed in detail. The CRM Model is characterized by transparent conversions between heterogeneous
data formats using MDI (metadata information), communication models from one to the other sources of
data using broadcasting, and united programming interfaces. This model of Data Mining conducts an
access-processing application. To study the effects of the proposed model, the case based on mining fact
rules are designed and implemented. The result shows that the models is feasible. This study’s
conclusions could indicate that provided a possible thoughtful method of integration on data mining of
stock customer relationship management.
Key words Customer Relationship Management(CRM), Data Mining, Business process Modeling
1 Introduction
Customers’ loyalty is the most valuable wealth of modern enterprises in era of knowledge economy.
Actually, management and marketing of an enterprise aim at turning potential customers into factual
customers, satisfied customers into loyal lifetime customers. A new management thought and technology,
Customer Relationship Management (CRM), came into being under the circumstances. The CRM
requires that the organizations tailor their products and services and interact with their customers based
on actual customer preferences, rather than some assumed general characteristics. Firms today are
concerned with increasing customer value through analysis of the customer lifecycle. The tools and
technologies of data warehousing[1], data mining and knowledge discovery [2][3][4][5]and other CRM
techniques afford new opportunities for businesses to act on the concepts of relationship marketing.
Through an in-depth analysis of customers’ detail information, CRM can acquire new customers,
reserve customers and increase customer profit contribution degree in order to obtain or maintain
competitive advantages. CRM can be viewed as ‘Managerial efforts to manage business interactions with
customers by combining business processes and technologies that seek to understand a company’s
customers[1]. Companies are becoming increasingly aware of the many potential benefits provided by
CRM. Some potential benefits of CRM are as follows: (1) Increased customer retention and loyalty, (2)
Higher customer profitability, (3) Creation value for the customer, (4) Customization of products and
services, (5) Lower process, higher quality products and services [6][7].
The old model of “design-build-sell”(a product-oriented view) is being replaced by
“sell-build-redesign” (a customer-oriented view). The traditional process of massmarketing is being
challenged by the new approach of one-to-one marketing[8][9]. In the traditional process, the marketing
goal is to reach more customers and expand the customer base. But given the high cost of acquiring new
customers, it makes better sense to conduct business with current customers. In so doing, the marketing
294
focus shifts away from the breadth of customer base to the depth of each customer’s needs. The
performance metric changes from market share to so-called “wallet share”. Businesses do not just deal
with customers in order to make transactions; they turn the opportunity to sell products into a service
experience and endeavor to establish a long-term relationship with each customer. Cap Gemini conducted
a study to gauge company awareness and preparation of a CRM strategy. Of the firms surveyed, 65%
were aware of CRM technology and methods; 28% had CRM projects under study or in the
implementation phase;12% were in the operational phase. Therefore, someone regard customers’
information as the most valuable resources in the 21st century[10][11][12].
The purpose of this paper aims at applying data warehouse, data mining, three-ply architecture and
Agent technology to construct a data mining model of CRM based on WEB data warehouse and analyze
a part of system distribution as an example. The result shows that the model is feasible.
2 Data Mining System Model of CRM Based on Web Data Warehouse
2.1 Data Mining for CRM of Architecture
The most valuable purpose of CRM system is to send correlative customers’ right information to a right
person at a right time in security and assist managers to make the best decision. Agent is a substitution for
people to search and manage correlation knowledge, which can accept consignment and process
perception, reaction, layout, modeling, communication and decisions, including abilities to negotiate,
cooperate and so on[13] .
Nowadays, people gradually realize that they are in a data trouble but hunger for knowledge. In order
to solve this problem, the technology of data warehouse and date mining came into being[14]. Data
warehouse supports the decision and management process. It is a theme-oriented and integrated data set
that cannot be upgraded but change along with time[15]. Data mining is a process to mine interesting
knowledge from mass data stored in database, data warehouse and other information base. Namely it
means to find the relationship and model among data elements.
According to the thought mentioned above, we present an assistant decision-making model of CRM
based on web data warehouse. The thoughtway is given in Fig. 1. Thereinto, CDMS: Case Database
Management System;KDMS: Knowledge Database Management System; MDMS:Model Database
Management System; ADAS: Method Database Management System;ODA: Method Database Agent ;
KDA: Knowledge Database Agent;MDA: Methed Database Agent ;ADA:Case Database Agent;
295
Time
dimension
Product
dimension
Seller dimension
Granularity
Fig. 2
The multi-dimensions data model of
sales management
Dimension
Time
Product
Seller
Year
Area
Name
State
Dept
Quarte
r
Month
Sort
Genre
Province
Group
Type
Quantit
y
…
City
…
Week
Region
Dept.
Quantit
…
y
Day
…
The fact: present period sales volume, present period
sale amount, accumulative total sales volume,
accumulative total sale amount and sales volume
fluctuate at same term.
Fig. 1. The CRM model for data mining of WEB
Fig. 3
The Conceptual model of sales
management
2.2 Modeling Methods of Data Mining
The problem exists in the disparity between the current situation and ideal and expecting situation. The
solution of problem removes this disparity[16]. Agent of problem converts and classifies problem asked by
users, matching that problem in database. If it is solved by computer solving system, a factual question
will become an example, otherwise knowledge information model of the problem will be added into the
database of problem case by management system of problem database.
Knowledge database comprises knowledge to solve problems such as various regulations and experts’
experience. It is the foundation that using methods match the model to deduce and calculate. To enhance
the systemic intelligence, knowledge base must be able to obtain new knowledge in the changing
environment and accumulate knowledge based on the original. Referring to this system, knowledge base
is a dynamic concept because it is added into new knowledge through continually data mining besides the
original knowledge. According to material question, Model Database Agent requires one model or
several models to match.
Management systems, such as CDMS, KDMS, MDMS and ADMS, add, delete, modify and query data
in each database separately, satisfy the needs of new conditions to keep the systems opening and
adaptable.
296
Database of approach consists of methods to solve problems, methods based on the models calculate
according to correlation methods that models choose[17]. The application of Method Databse enhances
adaptation of systems more. When encountering a problem, systems can choose one model but the
methods of solution are selective. Data warehouse contains data accumulated for a long time and it is the
source of the whole system. Therefore, data warehouse is the foundation stone of the assistant
decision-making model of CRM. However, data from warehouse are huge, system can not support the
time and space of calculation. Generally, agent processes in advance and makes a subset according to
specific problems.
3 Case of Analyses
3.1 The System Structure
We apply the model of knowledge structure mentioned above to CRM software of certain stock
enterprise and it obtains good results. Because the whole systems involve many aspects of CRM, we set
seller-management as an example to illustrate only.
In order to avoid high coupling between application system and specific platform, we introduce
middleware technology, which forms the B/S Browser/Server of three-level architecture based on web
browser, web server and application server.
The main tools used in our system are WEBSPHERE+DB2 of IBM and Prolog language. The
implementation mechanism is that user send request of data access to application logic using HTTP
(hypertext transfer protocol) protocol through user interface of WEB browser. In fact, B/S
Browser/Server can share application logic by all users , but C/S (Client/Server) model can not share
that. Then, the application logic server accesses the back-ground database by instancing the Javabeans
and save the results in Javabeans. Finally, JSP (JavaServer Pages) supported by WEBSPHERE processes
data obtained from calling Javabeans, and return to client in the way of dynamic HTML (HyperText
Markup Language) page. All the logics almost execute on the server, and require little for the client.
（
（
）
）
3.2 The Market Segmentation and Deduction of Demand Rule
According to the theory of market segmentations, enterprises must find the target market and give
products an exact position. Pi represents the number i market segmentations and Qi represents the
demand of market segmentations Pi. Thus, we define a set of user-demand regulation with IF-THEN
form.
Where the fuzzy implicate relationship of the i-th rule can be described as equation (1)
Pi → Q i , ( Pi ∈ P , Q i ∈ Q )
(1)
In other words, the demand is Qi if customers belong to market segmentations Pi.
If another market subdivision P’ is given, the demand Q’ can be derived as equation (2)
Q ' = (P → Q ) o P '
(2)
As is well known, evaluating customer profitability, the law of Pareto 20/80 means that 80% of the
profits are produced by top 20% of profitable customers and 80% of the costs are produced by top 20% of
unprofitable customers. The core parts of CRM activities are understanding customers’ profitability and
retain profitable customers. As a kind of strategic resources, to cultivate the full profit potentials of
customers, many companies already try to measure and use customer value in their management
activities[17].
One purpose of CRM aims at using data warehouse and correlation technique to find out 20 percent of
optima customers for seeking different competitive advantage with differentiation strategy. Therefore,
many firms are needed to assess their customers’ value and build strategies to retain profitable customers.
3.3 The Data Analyses
297
When designing and implementing sales and management of CRM, one basic problem encountered is
to ascertain, find and trace who makes the purchase and when, where and how. Obviously, it is a
multi-dimensional problem. We use hypercube structure to describe the fact of sales management,
dimension and granularity relation. As is shown in Fig. 2, each coordinate axes is one dimension, the unit
of coordinate axes is granularity and the fact represents each point in data space. The illustration is given
in Fig. 2. A certain point in data space expresses the quantity purchased of some suppliers at a certain
time.
It is described in formalization as follow: set P is defined as the aggregate of all enterprises and the
definition of D is the aggregate of potential sellers. And the Descartes product, V=P×D, constitutes entire
sale of all the products. R, product sales relation, is binary relation of the set, namely, product vs sales
<Va,Vx>. E represents the validity of relation and it can be divided into the validity of time (t) ,the validity
of products (p) and the validity of sale (d) and so on, then the validity of sale relation (T/F) is determined
by E,
E(t,p,d)=t p d
(3)
(4)
a
x
∧∧
R={<V ,V >|a∈P; x∈D }
R[ E (t , p, d ) = T ] 
R=R(E(t,p,d))= 

R [ E (t , p, d ) = F ]
(5)
When it is true, the formula above expresses that the relation of product sale exists, or does not exist.
On the basis of specific problem, we summarize the fact of product sale, dimension, granularity and
the relation among them. The conceptual model of sales management is given in Fig. 3.
From Fig. 3, we can obtain the star-type logical model of sales management of CRM, as the Fig. 4 is
shown.
Time dimension
T Prime key
Day
Month
Quarter
Year
…
Seller dimension
S Prime key
S Name
S Genre
S Grade
…
Product dimension
Fact of sales
T Prime key
P Prime key
S Prime key
A Prime key
Present
period sales
volume,
present
period sale
amount,
accumulativ
e total sales
volume …
P Prime key
P Name
P Grade
Volume
Weight
Price
…
Area dimension
A
Prime
key
State
Province
City
Fig. 4 The star-type logical model of sales management of CRM
4 Discussion
Technologies such as data warehousing, data mining, and campaign management software have
made customer relationship management a new area where firms can gain a competitive advantage.
298
Particularly through data mining—the extraction of hidden predictive information from large
databases—organizations can identify valuable customers, predict future behaviors, and enable firms to
make proactive, knowledge-driven decisions. The automated, future-oriented analyses made possible by
data mining move beyond the analyses of past events typically provided by history-oriented tools such as
decision support systems. Data mining tools answer business questions that in the past were too
time-consuming to pursue. Yet, it is the answers to these questions make customer relationship
management possible[19].
We present an integrated framework for knowledge management in the context of marketing, we
realize that there are critical research challenges to be addressed. another important challenge is a web
mining.With the Internet emerging as the new channel for distribution of goods, promotion of products,
handling of transactions, and coordination of business processes, the Web is emerging as an important
and convenient source of customer data. But, the multiple data formats and distributed nature of the
knowledge on the Web make it a challenge to collect, discover, organize and manage the knowledge in a
manner that is useful for marketing decision support. As marketing depends more and more on the Web
for customer data, Web mining need to be addressed as an important marketing knowledge management
problem.
5 Conclusion
The frame construction of system brought forward by this paper and it can assist a company to work out
certain forward-looking decision based on knowledge. Nevertheless, the data mining's process of stock
customer is the high complexity of intelligent problems for exploiting heterogeneous system, how to
identify, describe and analyze the process in the dynamic environment is the foundation of knowing and
improving CRM. Because of lack of the systematic theoretical guidance, systematic implementation
method and technology has not formed for the changes management of the process yet at present, and the
theoretical foundation of design and appraise method is in lack. The process modeling, simulation, and
implement are three important stages of variation management in enterprises, and they are needed to
integrate effectively. However, there are still some questions about knowledge induction and database
rewritten and refine.
References
[1] Usama M. Fayyad. Data Mining and Knowledge Discovery:Making Sense Out of Data. IEEE Expert:
Intelligent Systems and Their Applications, 1996, 11(5): 20~25.
[2] Ming-Syan Chen, Jiawei Han, Philip S. Yu. Data Mining: An Overview from a Database Perspective.
IEEE Transactions on Knowledge and Data Engineering, 1996, 8( 6):866~883.
[3] D.A. Keim. Visualization techniques for mining large databases: a comparison. IEEE Transactions
on Knowledge and Data Engineering. 1996, 8(6): 923~937.
[4] M.C. Rumizen. Report on the second comparative study of knowledge creation conference. Journal
of Knowledge Management, 1998, 2(1): 77–82.
[5] R. Agrawal, T. Dimielinski, A. Swami. Database mining: a performance perspective. IEEE
Transactions on Knowledge
and Data Engineering, 1993, 5(6): 914–925.
[6] Kim, J., Suh, E., & Hwang, H.. A model for evaluating the effectiveness of CRM using the balanced
scorecard. Journal of Interactive Marketing, 2003, 17(2):5–19.
[7] S.-Y. Kim, Tae-Soo Jung, Eui-Ho Suh, Hyun-Seok Hwang. Customer segmentation and strategy
development based on customer lifetime value: A case study. Expert Systems with Applications,
2006, 31(1):101–107.
299
[8] Stone, M., Woodcock, N., & Wilson, M.. Managing the change from marketing planning to customer
relationship management. Long Range Planning, 1996, 29(5): 675–683.
[9] D. Peppers, M. Rogers. Is your company ready for one-to-one marketing? Harvard Business Review,
1999, 77(1): 151–160.
[10]
Pa Hanaik P K, Sengupta K. On the structure of simple Preference-based choice functions.
Social Choice and Welfare, 2000, 17(1):35-40.
[11]
Dean A Shepherd,Moren Levesqne. A search strategy for assessing a business opportunity.
IEEE Transactions Engineering Management, 2002, 49( 1):140- 151.
[12]
Spreng R A, MacKenzie S B, Olshavsky R W. A reexamination of the determinants of consumer
satisfaction. Journal of Marketing, 1996,60(3):15-38.
[13]
WANG Zhong-tuo. Establishing a new discipline: Knowledge systems engineering, Journal of
Dalian University of Technology, 2000,40,s1(12):115-122(in Chinese).
[14]
HUANG Fei-xue et al. Technology of Data Mining Based on Neural Network, Application
Research of Computer, 2000,17(6):4-9(in Chinese).
[15]
W. H. Inmon. The Data Warehouse and Data Mining. Communication of the ACM,
1996,39(11):49-50.
[16]
WANG Zhong-tuo. Meta-decision-making: Concept and Methodology. Journal of Dalian
University of Technology(Social Sciences),1999,20(2):1-7(in Chinese).
[17]
Ta-Tao Chuang et al. The development of an adaptive decision support system. Decision
Support System, 1998 24(2):73-87.
[18]
Verhoef, P. C., & Donkers, B. Predicting customer potential value an application in the
insurance industry. Decision Support Systems, 2001, 32(2):189-199.
[19]
Chris Rygielski, Jyun-Cheng Wang, David C. Yen. Data mining techniques for customer
relationship management.
Technology in Society, 2002, 24(4): 483–502.
300