International Journal of Scientific & Engineering Research Volume 2, Issue 5, May-2011 1
ISSN 2229-5518
Adaptive Classification of Web Mining Methods and Challenges of Customer Relationship Management Domain
Siavash Emtiyaz, MohammadReza Keyvanpour
Abstract — In recent years, World W ide Web has been extended from research society to the most dominant and general way for communication and broadcasting of information. Web mining is responsible to discover the hidden knowledge, rules and patterns from web. Customers always play a key role for the establishment or mean of crisis for any organization. Web mining is going to be involved in every organization for extracting extra information which is not visible for everyone. The most important application of web mining is in the domain of the e-commerce and economy that leads to the detection of most facts and effective factors in the customer relationship management and efficient services to the customers through the behavior and communicating with system. CRM uses data mining (one of the elements of CRM) techniques to interact with customers. It is also used for web mining in web domain.Our Analysis provides a roadmap to guide creation concerning the adaptive classification based on web/data mining methods to solve challenges of CRM. For this, we comparison and analyze the application of web mining process to solution challenges of CRM dimensions.
Index Terms— CRM Challenges, Customer Relationship Management, Web Mining, Data Mining.
—————————— • ——————————
HE e-commerce Rapid development Caused to busi- ness advancement in the Internet. The CRM became the main part of the work when e-commerce turned
to customer-oriented from product-oriented and products and services were built and personalized according to customer’s recommendations. It was done because the internet-based technologies are the best framework for implementation of CRM [1]. One of the goals of CRM is close relationship between customers and vendors, which is the simplest form of e-commerce. CRM uses data min- ing (one of the elements of CRM) techniques to interact with customers. It is also used for web mining in web domain. Web mining is one of the domains of Data min- ing. Web mining refers to the techniques that automatical- ly retrieve, extract and evaluate information from docu- ments and web services for knowledge discovery. This article proposes data mining methods in the commerce, understanding challenges and ways to manage customer communications with the related rules and classification provided in this area.
CRM is collection of steps to create, develop, maintain and optimize long-term and valuable relationships between Customers and organizations. Furthermore, it is
————————————————
• Siavash Emtiyaz, Departments of Computer Engineering Qazvin Islamic
Azad University, Qazvin, Iran. E-mail: siavash1065@gmail.com
• MohammadReza Keyvanpour, Departments of Computer Engineering
Qazvin Islamic Azad University, Qazvin, Iran.
E-mail: keyvanpour@alzahra.ac.ir
part of the organization's strategy to identify customers, keep them satisfied and change them in to permanent customers. The main architecture includes several layers [3]. Fig (1)
Technologies used in the operational layer are responsible
for collecting customer data through their contact points
such as connection centers, connection management sys- tems, mails, faxes, sale persons, web and etc. Fig (2)
layers that are responsible to obtain, store, process, in-
terpret and report data to the users which use customer’s
data. This layer provides capabilities for classification of customers to optimize the behavior of company, improve
marketing activities and keeping customers. It is done by studying information in information storage. In terms of ICT and e-business development, customers' data are gathered more and more and day by day. Analytical layer capabilities can be used for better management of these data. Operational and analytical layers interact with each other. This means that data of operational layer go through to the analysis layer, then the data are analyzed and the results derived would have direct effect on the operational layer. By analyzing this section, Customers are classified and focusing on particular customers is provided for the organization.
IJSER © 2011 http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 2, Issue 5, May-2011 2
ISSN 2229-5518
This layer uses statistical analysis and predictive models for making decision, process improvement and other un- perceived benefits. Generally, they have been equipped by two types of technology: Online analytical processing, Data mining [3].
Online analytical processing is a process that uses one or more resources for collecting data. Data analysis and un- ification are done dynamically in different levels so they can help in finding, mapping and analyzing the custom- er's patterns based on their past data values.
customers' profiles are created, this layer provides some
statistics and analyzes the status of sales, marketing and support and it also measures the amount of satisfaction
and loyalty of customers then the customers will be cate- gorized based on this method. For example, we can give responsibility to the system for monitoring all of custom- er's behaviors and when the amount of customers’ pur- chases exceeds of a certain limit, makes the organization informed to do special works for them, Such as sending email or phone call automatically with the buyer.
Fig 1: CRM Architecture [3]
Marketing
Contact
Organization before implementing a CRM should be aware of the potential and possible problems, in order to deal with them when it is necessary.
In organizational levels, business should create common activity between different parts that have contacts with customer and all sections related to the CRM in order to be more effective.
This established inter-group relationships with custom-
ers, have effect on the role of employee within the
organization also this issue can reduce emphasis on par- ticipation of the certain sections of the company. Usually
this type of policy changes faced with the opposition in the early stages of performances.
One of the important points is that organizations should be maintaining the best employees of sales and service departments; they must develop personal skills of ta- lented employees and reward them. For example, in tele- phone centers, six months is needed to turn an armature person to an experienced one and six month is needed for an experienced person to become specialized and profes- sional. Employees, who work in the internal sections, should be involved in the discussions related to the cus- tomers' needs. Customers' feedbacks should be consi- dered in the process of developing products and servic- es. Because all parts of organization collaborate in order to gain customer satisfaction, they should create a reward system for injecting the necessary motivation to these domains. The main challenges can be divided into four important cases that organization might face with during the implementing of CRM [2, 6, 7]:
• Initial launching Costs: It is considered as one of the challenges of CRM. Organizations may spend large amount of investment on customer man- agement application tools, but some of these tools may have a specific application that can be hard to share them in different parts.
• Integrated application tools: Organizations need some integrated application tools that have been created based on customer's life cycles and re- lated interactions. For example, organizations which need different languages and monetary units for managing customer interactions, cannot
Customer
Data
Sales
Service
Media
Integration
implement CRM through traditional technolo- gies and this will be a severe problem, because data types are various and the noise of data is high and inevitable.
• Collaboration of various segments: CRM is an in- tegrated approach and needs assisting of some parts of business that previously were working independently. Data that are collected in one portion should to share their data with be shared within all segments. It is possible that some sec- tions would be unpleased others.
Challenges that should be noticed in this approach are:
Fig 2: Operation Layer CRM
IJSER © 2011 http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 2, Issue 5, May-2011 3
ISSN 2229-5518
1- Non-obvious results that often require a combination of the analysis layer techniques (data mining), 2- Serious requests for collecting of data before using of analysis layer techniques (data mining), 3- Not being aware that what data are available for mining and what kinds of virtual activities should be done.
These are parts of inter-organization challenges which provide forces that any decision should be made is based on it and Organization processes are managed from the starting point to the end within the rules of the Organization. The raised Challenges in this regard are:
1- output results must match the realities of the world,
2- good activities mechanism, 3- combination of the old knowledge.
These are one of the key challenges for the realization of policies, processes, people and their interaction with each other. Organizations need an integrated approach to uni- fy the technologies within it. The raised Challenges in this regard are: 1- Reliable considerations don't exist for re- sults of analysis and data because of lack of integrated approach in technologies. 2- Finding data for deeper un- derstanding, which cause high accuracy and cost reduc- tion, for example metric product evaluation.
These Challenges that relate to the customer loyalty and maintenance are: 1- Deeper models for identifying and developing customer behavior. 2- A framework for eva- luating that leads to diagnosing the accuracy of customer understanding.
The research method can be divided into three steps: (1) collecting and cleaning the information, store the in- formation in the data warehouse.
(2) An iterative discovering process by the data mining tools and analysts review of the extracted patterns to gen- erate new set of questions to refine the search. After refine the search, the results of the mining process is be trans- lated to association rules, which stored in the knowledge base.
(3) The patterns are good predictors of purchasing beha- viors, the CRM process uses the scores generated by the data mining process to sharpen the focus of targeted cus- tomers or prospects, thereby increasing response rates and campaign effectiveness. Fig (3) presents the applica- ble methods of web mining for CRM [1,9,10].
A. Data Collection Process
Suitable representation, in such a way that it is readable
by the mining algorithm in the next step. This process
ensures the data set is ready to use for specific mining technique and purpose. In this process, generalization can
be done through attribute removal or attribute generaliza-
tion and aggregation can be done by combines conformed records and add up total amount.
B. Data mining process
The Fig (4) process model for data mining provides an overview of the life cycle of a data mining project. It con- tains the corresponding phases of a project, their respec-
tive tasks, and relationships between these tasks. At this description level, it is not possible to identify all relation- ships. There possibly exist relationships between all data mining tasks depending on goals, background and inter- est of the user, and most importantly depending on the data.
Fig 4: The CRISP-DM process model [8]
C. CRM process
The CRM process is the most influential customer
oriented strategy of the decade. Despite its humble ori- gins it has evolved into a relatively complex strategy. The
essentials of a CRM program include focus, commitment to CRM goals and above all a desire to be customer fo- cused [8].
Data warehouse, data mining, used to serve the purpose
of supporting selection for acquisition, cross-sales, and
retention of (real or potential) customers. By running data mining algorithms on customer dataset, data mining can
uncover important associations about what products are often purchased together. This knowledge can then be used for product recommendations and product bun- dling. This knowledge is then used to make a recommen- dation for a future customer.
The most important results obtained from the above techniques are reorganization of data and information for easier access, yielding more effective Efficiency and better data classification in order to obtain best results from them. Web mining and CRM integration has advantages and lots of benefits for companies that need to discover profitability of some customers than other customers.
IJSER © 2011 http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 2, Issue 5, May-2011 4
ISSN 2229-5518
Web mining can identify important customers in large databases.
Intelligent CRM system is created based on web mining analysis on important customers that leads to customer's management ,sharing the obtained information from dif- ferent channels, reforming of communications between different parts, organization and investigation of opera- tional activities and the appropriate understanding of trade. Therefore, customer chooses the connection chan- nel with the company according to his or her own interest to receive the best services .In table (1) Proposed Adap- tive Classification Based On Web/Data Mining Methods to Solve Challenges of CRM are shown.
[1] Shi-Ming Huang, Irene Kwan and Shing-Han Li; "Web Mining for CRM an Empirical Study of computer Game Service Com- pany", International Journal of Electronic Business Manage- ment, Vol. 1, No. 1, pp. 36-45 (2003).
[2] Jaideep Srivastava;"Data Mining for Customer Relationship
Management ",white paper.
[3] Dezhen Feng, Zaimei Zhang, Fang Zhou, Jianheng Ji;" Applica- tion Study of Data Mining on Customer Relationship Manage- ment in E-Commerce"; 978-1-4244-3291-2/08/$25.00 ©2008
IEEE.
[4] Olaf Boon, Brian Corbitt, Craig Parker;"Conceptualizing the Requirements of CRM from an Organizational Prespective a Review of the Literature";School of Information Systems Deakin University,2002.
[5] E.W.T. Ngai , Li Xiu , D.C.K. Chau; "Application of data mining
techniques in customer relationship management: A literature review and classification"; Expert Systems with Applications journal, 2009 Elsevier.R. Nicole, "The Last Word on Decision Theory," J. Computer Vision, submitted for publication. (Pend- ing publication)
[6] Research Team CRMnext Knowledge Base;"CRM Challenges
Building an Effective Strategy"; directory group;
[7] Mirela Danubianu, Stanica Anca Maria;"Study Of Improving The Customer Relationship Management By Data Mining Ap- plication"; a lecturer in Electrical Engineering and Computer Science, 2009.S.P. Bingulac, “On the Compatibility of Adaptive Controllers,” Proc. Fourth Ann. Allerton Conf. Circuits and Systems Theory, pp. 8-16, 1994. (Conference proceedings).
[8] A.S. Al- Mudimigh, F. Saleem, Z. Ullah, F.N. Al-Aboud; "Im-
plementation of Data Mining Engine On CRM Improve Cus- tomer Satisfaction"; King Saud University, Riyadh, Kingdom of Saudi Arabia; 978-1-4244-4609-4/09/$25.00 ©2009 IEEE.
[9] R. Kosala,H. Blockeel; "Web Mining Research : A Survey", In
SIGKDD Explorations, ACM SIGKDD,Volume2,Issue1, July
2000
[10] Zdravko Markov, Daniel T. LAROSE;" Data Mining The Web Uncovering Patterns in Web Content, Structure, and Usage", WILEY-INTERSCIENCE, 2007.
[11] Jaideep Srivastava, Robert Cooley, Mukund Deshpande, Pang- Ning Tan; "Web usage mining: discovery And applications of usage patterns from web data".SIGKDD Explorations, 1(2):12–
23, 2000.
Data Collecting Process
Data Mining Process
CRM Process
Data Source Database, Flat Files
Operation CRM
Preprocess
Data Extraction Transformation Load
Discovering
Patterns
Refine
Patterns
Review
Patterns
Marketing Sales
Service
Data Warehouse
Knowledge
Fig 3: Presents the Applicable Methods of Web Mining for CRM [1]
IJSER © 2011 http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 2, Issue 5, May-2011 5
ISSN 2229-5518
Table 1. Proposed Adaptive Classification Based On Web/ Data Mining Methods to Solve Challenges of CRM
Mining Types of Methods Main Idea Dimensions of CRM CRM Challenges
Data mining
Association Rules
Classification Clustering Forecasting
Regression
Sequence
Visualization
Statistical and former algorithms are used for diagnosing the behavior of an event or specific process that identifies the communication between elements in a specified record uses. Classification is one of the most common learning models in data mining. It aims at building a model to predict future customer behaviors through classifying
Database records into a number of predefined classes based
on certain criteria. Common tools used for classification are
neural networks, decision trees and if then-else rules. Clustering is the task of segmenting a heterogeneous popu- lation into a number of more homogenous clusters.
It is different to classification in that clusters are unknown at
the time the algorithm starts. In other words, there are no predefined clusters. Common tools for clustering include
neural networks and discrimination analysis
Forecasting estimates the future value based on a record’s patterns. It deals with continuously valued outcomes. It relates to modeling and the logical relationships of the mod- el at some time in the future. Demand forecast is a typical example of a forecasting model. Common tools for forecast- ing include neural networks and survival analysis.
Regression is a kind of statistical estimation technique used to map each data object to a real value provide prediction value. Uses of regression include curve fitting, prediction (including forecasting), modeling of causal relationships, and testing scientific hypotheses about relationships be- tween variables. Common tools for regression include linear regression and logistic regression.
Sequence discovery is the identification of associations or patterns over time. Its goal is to model the states of the process generating the sequence or to extract and report deviation and trends over time. Common tools for sequence discovery are statistics and set theory.
Visualization refers to the presentation of data so that users can view complex It is used in conjunction with other data mining models to provide a clearer understanding of the discovered patterns or relationships. Examples of visualiza- tion model are 3D graphs.
Customer Development Integrated application tools
Technology challenges
Collaboration of various segments
Customer Identification Initial launching Costs
Strategy challenges
Customer Dimensions Challenges
Customer Identification Initial launching Costs
Strategy challenges
Customer Dimensions Challenges
Customer Attractions Integrated application tools
Collaboration of various segments
Customer Retention Collaboration of various segments
Strategy challenges
Customer Dimensions challenges
Customer Retention Collaboration of various segments
Strategy challenges
Technology challenges
Customer Development Integrated application tools
Technology challenges
Collaboration of various segments
Web
Web Content Mining
Active methods of Automatic discovery, retrieve, organiza- tion, and management of a huge volume of web information and resources. It helps in improving or filtering the search information that is usually based on the information derived from user profile or request so provides more complex queries from a simple keyword for more accuracy.
Customer Identification (Customer's Primary Infor- mation)
Initial launching Costs
Strategy challenges
Customer Dimensions Challenges
Mining
Web Structure Mining Web is a graph. Its nodes are web pages and hyperlinks create edges that provide the communication between the related pages.
Customer Identification (Customer's Primary Infor- mation)
Initial launching Costs
Strategy challenges
Customer Dimensions Challenges
Web Usage Mining
It focuses on techniques that can predict user behavior inte- ractively while working with the web pages. Mostly, e- commerce is very important for web based companies and CRM can be an effective advantage of web usage mining.
Customer Attractions, Re- tention, Development (Customer's Secondary In- formation)
Integrated application tools Technology challenges Collaboration of various segments Customer Dimensions Challenges
IJSER © 2011 http://www.ijser.org