Widespread changes within business environments in recent years has demanded acquisitions of new tools that are more skilled to cope with new challenges and demands in business. Advances in computer technologies, higher accessibility of computer associated tools and decreased prices of general computer-related products are reasons enough for at least considerations about higher usage of new technologies. Particularly in direct marketing activities discussed technology is called Data Mining.
Companies are faced with hosts of data collected in their data repositories. Of course, companies want to make use of their data and aim to discover interesting patterns of knowledge within their data repositories. Direct marketers which can be involved in catalogue marketing, telemarketing or widely known direct-mail marketing are intensive users of Data Mining Technologies. Because of that, the authors strive to do research concerning reasons for and advantages and disadvantages with using Data Mining as support for direct marketing activities.
Chapter 1 deals with general information for the reader as support for delving into the topic. The included problem discussion finishes with the final problem formulation of this thesis. Chapter 2 is about the Methodology which includes considerations of Gummesson. The following theoretical part is divided into two major parts, Data Mining and Direct Marketing, which underpin the whole thesis. The authors want to inform the reader about important and sophisticated contents concerning both Data Mining and Direct Marketing. Without overloading the implementations about Data Mining and Direct Marketing, the authors conduct the reader to essential and detailed aspects of both fields for understanding the intentions.
The empirical part contains a short introduction to each company within the thesis, and short summaries of the interviews conducted. In the following analysis part the authors have created a model to make the possible uses of Data Mining more understandable to the reader. Furthermore, this part contains an analysis of the interviews in relation to the topic at hand and the theories used. In the conclusions the authors answer their research question, namely; what are the main advantages and disadvantages of Data Mining as support to Direct Marketing activities? In the absolute end of the thesis the authors criticise their own work and give suggestions for further research within the topic.
Table of Contents
1.Introduction
1.1 Background
1.2 Problem Discussion
1.3 Problem Formulation
1.4 Purpose of this Thesis
1.5 Limitations
1.6 Theoretical Relevance
1.7 Practical Relevance
2. Methodology
2.1 Preunderstanding
2.2 Research Journey
2.3 Paradigms
2.4 The subjective - objective dimension
2.4.1 Ontology: Nominalism vs. Realism
2.4.2 Epistemology: Anti-positivism vs. Positivism
2.4.3 Human Nature: Voluntarism versus Determinism
2.4.4 Methodology: Ideographic versus Nomothetic
2.4.5 Radical change- Regulation
2.4.6 Scientific approach
3. Theory
3.1. Data Mining
3.1.1 Data Mining Methods
3.1.1.1 Data Clustering
3.1.1.2 Classification
3.1.2 Modes of Data Mining
3.1.2.1 Descriptive Data Mining
3.1.2.2 Predictive Data Mining
3.1.3 Overview of Data Mining Techniques
3.1.3.1 Market Basket Analysis / Association Mining
3.1.3.2 Artificial Neuronal Nets
3.1.4 Traditional Statistical Approach
3.1.5 Data Warehouse
3.2. Direct Marketing
3.2.1 Benefits and Growth of Direct Marketing
3.2.2 Customer Databases
3.2.2.1 Database Marketing
3.2.2.2 Mass Marketing versus One-to-One Marketing
3.2.3 Major Channels of Direct Marketing
3.2.3.1 Direct Mail
3.2.3.2 Catalogue Marketing
3.2.3.3 Telemarketing
3.2.4 Ethical Issues in Direct Marketing
4. Empirical Part
4.1 Telia
4.2 Skandia
4.3 SEB (Skandinaviska Enskilda Banken)
4.4 TUI Deutschland GmbH
4.5 Kreissparkasse Grafschaft Diepholz
4.6 OLB (Oldenburgische Landesbank AG)
5. Analysis
5.1 Usage of Data Mining today
5.2 Further Development of Data Mining
5.3 Main Advantages of Data Mining
5.4 Main Disadvantages of Data Mining
6. Conclusion
6.1 Critics to own work
6.2 Suggestions for further research
Acknowledgements
We are very grateful to the persons who have supported our master thesis in any way. On this track, many people helped us reach the final aim. People as our Tutor and Examinator, Mr. Stefan Lagrosen, all Interviewees, who have spent time answering our interviews and giving us valuable material to work on, made it possible to conduct our thesis.
We want to thank especially Mrs. Karin Guselin, Mr. Marcus Bove, Mr. Dennis Ekroth, Mr. Torsten Gebke, Mr. Rolf Vielhauer, Mr. Mathias Kording and Mr. Uwe Noster for answering our questions.Our thank is also pointed to people of our personal environment giving us help and support all time.
Växjö, January 2004
Tobias Brüggemann Patrik Hedström Martin Josefsson
Abstract
Widespread changes within business environments in recent years has demanded acquisitions of new tools that are more skilled to cope with new challenges and demands in business. Advances in computer technologies, higher accessibility of computer associated tools and decreased prices of general computer-related products are reasons enough for at least considerations about higher usage of new technologies. Particularly in direct marketing activities discussed technology is called Data Mining.
Companies are faced with hosts of data collected in their data repositories. Of course, companies want to make use of their data and aim to discover interesting patterns of knowledge within their data repositories. Direct marketers which can be involved in catalogue marketing, telemarketing or widely known direct-mail marketing are intensive users of Data Mining Technologies. Because of that, the authors strive to do research concerning reasons for and advantages and disadvantages with using Data Mining as support for direct marketing activities.
Chapter 1 deals with general information for the reader as support for delving into the topic. The included problem discussion finishes with the final problem formulation of this thesis. Chapter 2 is about the Methodology which includes considerations of Gummesson. The following theoretical part is divided into two major parts, Data Mining and Direct Marketing, which underpin the whole thesis. The authors want to inform the reader about important and sophisticated contents concerning both Data Mining and Direct Marketing. Without overloading the implementations about Data Mining and Direct Marketing, the authors conduct the reader to essential and detailed aspects of both fields for understanding the intentions.
The empirical part contains a short introduction to each company within the thesis, and short summaries of the interviews conducted. In the following analysis part the authors have created a model to make the possible uses of Data Mining more understandable to the reader. Furthermore, this part contains an analysis of the interviews in relation to the topic at hand and the theories used. In the conclusions the authors answer their research question, namely; what are the main advantages and disadvantages of Data Mining as support to Direct Marketing activities? In the absolute end of the thesis the authors criticise their own work and give suggestions for further research within the topic.
1. Introduction
When discussing a topic related to the field of information-technology it is very sensible to explain in a short manner what Data Mining is about: Data Mining is a process of discovering interesting knowledge. The sources of the knowledge, the data, are based in data warehouses, databases or other information repositories. The vast amount of data is collected by using enablers like the internet, bonus-cards or bar-code reader as some examples for gathering data. The aim is to extract ultimately understandable knowledge from huge datasets, which is interesting, useful and previously unknown.1
Knowledge can be depicted as identified patterns, significant structures or associations within the host of data. Data Mining can be applied to numerous areas such as banking and finance, retail or marketing. Used in marketing, Data Mining can be deployed within database marketing systems. The obvious purpose is to enhance the success rate of marketing activities.2
The applications of Data Mining (DM) Techniques to database marketing offer a broad range of opportunities. Analysis of supermarket point-of-sale data or predicting who the audience will be for television programmes are examples for the use and the impact of DM Techniques to marketing. Managers are faced a rapidly evolving competitive environment. Therefore their focus is shifting away from what products to sell to which customers to target. Targeting existing products to the right customer is a description for Database Marketing and this type of marketing segmentation used by businesses through knowledge discovery. It deals with direct marketing efforts either. The deployment of the tools is also focused to predict consumers´ behaviour and to identify these consumers being the most responsive to promotional and sales campaigns.3
1.1 Background
Studies made by Consulting Companies often claim that DM Technologies offer various opportunities for an increasing part of the companies in various branches. Particularly, the studies highlight the improved possibilities to achieve more and essential knowledge, which was not revealed before. By using methods as machine learning, neuronal nets or statistics, the DM searches for patterns, which need to be analyzed and interpreted to identify the new knowledge.4
With regard to the association of DM and Database Marketing, companies often possess a host of information about their own customers and interesting potential customers. Detailed data concerning demographic relations, communication-data, buying-information or potential- data are available.
In general it is not a real problem to get answers to quite low-demanding questions such as the average age of all customers. But often companies do not make use of their databases to get answers to real high decisive and essential questions because they do not have the necessary technologies and specialised employees. Examples of important questions are the extent of the danger of loosing customers or the possible turnover per customer in the next year. Other quite interesting and important questions could concern to the lifetime-profit per customer, the cross-selling potential of each product or just the simple but highly important question regarding the offer of the right product at the right time to each customer.5
To be able to answer the latter possible questions related to the data-based marketing, the correct combination of customer information is relevant rather than some listed characteristics of the customers. For example, the probability of a customer buying a certain product could depend on several factors such as age, gender, demographic typologies, the interest shown to other related or not related products, financial status or just being non-vegetarian. By combining the identified factors, describing interesting new or existing customers, DM technologies help to realize that.6
Other technologies such as statistical methods are common in Database Marketing, but they are not able to produce the same results as DM Technologies. The DM Technologies benefit from the advances of computer technology and related networks in general. By using artificial intelligence, famous examples are neuronal networks or rule-based systems, DM
Technologies can reach remarkable differences in comparison to other traditional methods such as statistical methods.7
1.2 Problem Discussion
The usage of DM is one opportunity to cope with the increased competitive pressure in business. From a commercial point of view the process of DM is very appealing to apply in marketing practices. Decisions need to be made rapidly and business in general is faced with an increased awareness of the need to know their customers. Therefore knowledge is essential, and a responsible treatment of sources for gathering knowledge is one major competitive advantage. In order to enable that, DM offers methods focusing on that.8
The deployment of DM Technologies in the area of marketing often leads to considerations regarding Direct Marketing efforts. Direct Marketing is the counterpart of mass marketing. Mass Marketing is mass media promotion; for example television advertising. In contrast, Direct Marketing tries to address consumers more precisely, accordingly to the consumers’ personal preferences and taste. The idea is to get to know each customer well. Thus the marketer can better serve each customer and increase the number of sales.
These considerations are very similar to the local merchant of the old neighbourhood store. The merchant knew you and your family, your preferences and tastes and built up a one-to- one relationship. Customers appreciated the friendly and personal atmosphere. Shopkeepers knew the names of the customers and recommended items which were useful to their customers. Consequently, the local merchant was able to keep a major part of the customers as loyal customers.9
It is obvious what the direct marketing pursues, namely a market segment of one person. The mass marketing is more a “big black box”, which cannot deliver enough information to know each customer. Furthermore it is not able to address the customers in the individualized fashion that direct marketing is.
A great American grocery chain, Dorothy Lane, discovered that 30% of regular customers account for 82% of its sales by using a loyalty card. In return the company pampers their identified important customers by gifts and discounts.10 Individualized catalogues by large retailers for different groups of consumers, loyalty programmes run by airlines or specialized mails to potential good customers are practical results of direct marketing activities.11 Direct marketing can benefit extremely from the improvements of sophisticated marketing database technologies.12 This also means that sophisticated marketing database technologies can, but do not have to, ensure improvements of marketing activities and their success rates. According to those conditions, the differences in the results when using DM technologies in comparison to other tools are very interesting.
Since it often is claimed that DM is such an advanced tool for developed Direct Marketing activities, it is worth to focus on this issue and to clarify the main advantages and disadvantages in using DM concerning Direct Marketing activities. It is necessary to delve into the depth. Obvious advantages, as shown before, are also faced with drawbacks. One example is that a lot of the people in deciding positions in companies are reluctant towards change, especially technological change. On the other hand new and advanced technologies are often seen as the solution for all unsolved problems. The deployment of new technologies often cannot fulfil the highly sophisticated expectations. This contradictory situation it is necessary to clarify key advantages and disadvantages of using DM as another relatively new technology.
1.3 Problem Formulation
Generally a company is faced with the question of how the data can help in solving their problems. This means that the DM technologies can help to define the problem. A general overview about existing problems and possible solutions often exist, but not the exact definition of the problem. The host of data in companies’ repositories sometimes reveals interesting information.13 With the support of new technologies such as DM, companies are able to identify problems, define them and find ways of solving them.
According to this, the topic of this thesis is not the possible usage of DM technologies, but the main advantages and disadvantages. This thesis focus lies on DM and it’s usefulness within
Direct Marketing. More precisely, it deals with direct marketing activities that are supported by sophisticated data based technologies such as DM. This leads the authors to the question;
What are the main advantages and disadvantages in using DM technologies as support to direct marketing activities?
1.4 Purpose of this thesis
The authors strive to find the main advantages and disadvantages of the usage of DM as support to direct marketing activities. They will attempt to highlight the reasons of the deployment of DM within the data based marketing, emphasizing direct marketing activities. The authors will try to find out if DM is a technique which is helpful for companies to use within the targeting for their Direct Marketing efforts.
1.5 Limitations
The focus of this thesis is addressed to interested marketing researchers rather than to DMexperts. Nevertheless, it is mandatory to gain basic knowledge about the DM tools that generate the useful data for selected marketing activities. Since the authors lack the computer science knowledge to go deep into the DM techniques the thesis is aimed towards the marketing activities and not the development of the DM techniques.
1.6 Theoretical Relevance
This thesis aims to contribute to a theoretical framework about the usage of DM within the targeting for Direct Marketing. The authors will try to investigate if there are advantages with using DM compared to other methods of targeting, or if the disadvantages supersede.
1.7 Practical Relevance
DM is a concept for accessing, searching and manipulating data to produce information that is useful to an organization. Relationships and meaningful trends can be predicted and increased success rates of sales campaigns are possible. Especially useful for direct marketing activities, deployed DM techniques can enhance the response of such activities as telephone calls, personalized mails or specialized catalogues.
2. Methodology
2.1 Preunderstanding
All people have different insights into different problems and their social environment; this is what is called preunderstanding.14 If the researcher does not have a basic preunderstanding on the topic he will study he will have to spend a lot of time on collecting basic information. On top of the previous knowledge, the researchers’ commitment to the topic is also included in the preunderstanding.15
Gummeson separates between two different types of preunderstanding; first he has the first- hand preunderstanding gained through personal experience. Secondly there is second-hand preunderstanding, gained through intermediaries e.g. lectures, textbooks and research reports.16
The authors of this thesis have different preunderstanding based on their different academic backgrounds. The two Swedish authors have gained their preunderstanding at Växjö University studying business economics. The third author with a deeper background of business informatics has studied business economics in Germany. From these perspectives the authors started their research on the topic at hand.
2.2 Research Journey
The journey in to the subject of Direct Marketing with the help of DM started with forming a study group consisting of two Swedish and one German student. The three had studied together during fall semester 2003 and they felt that they ought to be able to reach a good result in writing a thesis together. Their first meeting was a brainstorming session where they discussed what topics where interesting for them to write a thesis on. They finally agreed on the topic of Direct Marketing and DM. When the topic was chosen they tried to decide within what business areas they would try to contact companies. They decided that the financial sector, and especially credit cards companies, must be interested in the possible advantages of DM when trying to recruit new customers. When the process got further along the way the authors realized that it might be better to broaden their choice of companies and business areas, they also contacted companies within the regular banking sector, the telecom sector, the insurance business and travelling agencies. All these different branches of business could be helped by a better targeting for their direct marketing since they have products that are group specific, meaning that the customer needs to be a part of a certain group to be allowed to buy the product.
The thesis work went on with the formulation of a problem and a research question, this research question has been under constant development throughout the whole thesis writing period. The final question was:
What are the main advantages and disadvantages in using DM technologies as support to direct marketing activities?
To find the answer to this question the authors delved deep into the theories concerning Data Mining and Direct Marketing. From their newfound theoretical knowledge the authors created a questionnaire in which they decided what topics to discuss in their interviews in order to get a more practical understanding of their chosen topic. Instead of putting together a questionnaire with the precise questions they decided to focus on certain areas and to keep an open dialogue with starting point in these during the interviews.
The authors finally found a few good companies whose marketing departments they could use to find the information they were looking for. These companies were, Skandia, SEB, TUI, Telia, OLB and Kreissparkasse Diepholz.
The interviews were conducted in different manners some were over telephone and some were in a personal meeting with the interviewee. By analysing the results of these interviews the authors finally came to their conclusions.
2.3 Paradigms
The paradigm of a researcher is the frame of reference which governs the researcher’s instincts. The paradigm refers to the researcher’s insight, value judgement and standards.17 It is the theoretical and philosophical framework of a scientific school or discipline within which theories and generalizations and the experiments performed in support of them are formulated.18 Therefore the researcher must ask himself the question on which methodological approach he should use in his research.
Gummeson claims that there are two different approaches. Either the researcher uses the positivistic approach which wants to use certain cause and effect to explain relations; this is called the natural school. Or he could use the hermeneutic approach which tries to interpret situations in order to understand the observations conducted rather than to explain them.19
In the writing of this thesis the authors will try to be positivistic, they will try to find the effect that can be correlated to the usage of Data Mining in the targeting for Direct Marketing.
2.4 The subjective - objective dimension
There is a framework for research work within the social science area provided by Burrell & Morgan. In this they claim that there are two dimensions, the dimension of subjectivity and objectivity, and the dimension of radical change and regulation. These dimensions are described in the following scheme:
Abbildung in dieser Leseprobe nicht enthalten
Figure 1: Burrell & Morgan (1992), p. 22
The scheme is divided in to four quadrants, each representing an extreme appearance of intellectual thought. Burrell & Morgan do not exclude the possibility that researchers can allocate their own paradigm somewhere in the middle, for example a person can be allocated between functionalist and interpretative.20
Burrell & Morgan also developed a systematic model to have a closer look at the dimension about the subjective/objective debate in social science.
Abbildung in dieser Leseprobe nicht enthalten
Figure 2: Burrell & Morgan (1992), p.3
2.4.1 Ontology: Nominalism vs. Realism
The ontological sub-dimension relates to what is investigated, Nomalism is the subjective position; it implies that there is no real structure of the world. Thus the names, labels, and concepts that are used by humans to explain reality is nothing else than an easy way for people to understand the world in which they live. Realism is built on concrete, hard, and partly tangible structures. The opinion of a realist is that the world is independent from the individual and his perception. A person can do little or nothing to change or shape it.21
Here the authors will try to be realistic in their approach, and try to find the hard facts on the advantages and disadvantages with the usage of Data Mining for Direct Marketing purposes.
2.4.2 Epistemology: Anti-positivism vs. Positivism
Epistemology is the question of how the researcher would understand the data and how he would communicate those to his readers. It refers to the nature, origin and limits of human knowledge, how it is generated and it particularly concerns scientific knowledge. The subjective side claims that the anti-positivistic approach is better, they say that only people who are directly involved in the activities should be integrated in the study. According to anti- positivism these persons are the only ones that truly understands what is investigated, and this is what the anti-positivistic researchers aim to understand, the human activities. The positivism, on the other hand, offers an objective point of view. The positivist ones consider the knowledge to be hard, real and capable for transmission in tangible form. Positivistic researchers try to explain what happens through searching for regularities and causations of the studied objects. Thus, new insights might be made to existing knowledge and a hypothesis might be once again verified or finally rejected.22
Here the authors of this thesis will try to have an positivistic approach since they feel that their work should be understandable and useful for people in all parts of companies and society not only to those who are professional marketers.
2.4.3 Human Nature: Voluntarism versus Determinism
The human nature sub-dimension treats the relationship between the people and the environment in which they live. The voluntaristic approach claims that the humans can create their own environment. They also claim that individuals are free-willed and could exercise their ability to influence the environment in which they act. The deterministic approach tends to see human being as a product of her environment; humans are conditioned by their external circumstances. As a result it regards the environment as having control over the human’s free will.
2.4.4 Methodology: Ideographic versus Nomothetic
Methodology is the question of which kinds of methods we utilize to get information and the way to draw conclusions and how we build up the whole research and investigation process on our subject. The ideographic view is based on the point that one can only understand the social world by obtaining first hand knowledge of the subject under investigation. According to this view, the researcher is enabled to understand the subject achieving an inside view. The nomothetic view emphasises the importance of basing the research on systematic protocols and techniques. It offers methods which are grounded in natural sciences for example the construction of tests and the use of quantative techniques for the collection of data.23
Here the authors are Ideographic in their approach, they realize that it is not possible to have an opinion on something they are not familiar with, and therefore they will try to gain first hand knowledge within their chosen topic.
2.4.5 Radical change- Regulation
The second dimension in Burrell & Morgan’s model is the dimension of radical change and sociology of regulations. Radical change and regulation are closely related to political environment and as a result could lead to more controversial topics than the subjective or objective dimension. For example the sociology of radical change takes the ideas of Marx and resembles anarchistic individualism. Meanwhile the sociology of regulation includes ideas related to the epistemology of positivism. Supporters of radical change could be seen as a force behind rebellion and conflict, on the other hand supporters and regulation would try to find a way to keep society together and promote stability.24
2.4.6 Scientific approach
The scientific approach is the way the empirical research is performed. The researchers approach to reality will have an impact on the whole study and its result. There are two opposite approaches the researcher can chose, deduction and induction. In the deduction the researcher takes his starting point in existing theory or concepts and formulates hypotheses that will be tested, the profit of this is received theory. In the induction the researcher takes his starting point in real data, categories and concepts or patterns and models, this subsequently leads to the creating of theory. It is actually mostly the starting point that is affected by the choice of approach, but the distinction is essential. Deduction is used to test existing theories and induction is used to develop new theories. When the researcher has started his research, the process becomes an iteration of the two approaches; this is sometimes referred to as abductive research.25
The purpose of research can be divided in three areas: the exploratory, the descriptive, and the explanatory research. The descriptive approach gives answers to the following questions: what, where, when, and how; the researcher observes and describes situations and events. Researchers usually go on to examine why the observed patterns exist and what these patterns imply. This general purpose of scientific research is called explanatory research. Both approaches mentioned above are applied to established areas of research where the topic is clearly outlined. In cases where the researcher examines a field which is somewhat new and he wants to increase his knowledge, the exploratory research design is chosen. It is typically used for three purposes: (1) to satisfy the researcher’s curiosity and desire for better understanding, (2) to test the feasibility of undertaking a more extensive study, and (3) to develop the methods to be utilized in any subsequent study.26
Since the area of choice in this thesis is relatively new the authors will exploratory design which is closely related to the inductive approach.
3. Theory
The theory-chapter is divided into two major parts; the backgrounds of DM and Direct Marketing. The authors strive to emphasise the linkages between both concepts that underpin the problem formulation of the thesis. The theories of Philip Kotler, an US-American marketing professor, mainly underlay the part of Direct Marketing. The considerations of Usama Fayyad, Michael J. Berry, Gordon Linoff and George Marakas are the basic for the implementations within the DM part. In chapter three the authors will try to highlight the key parts of DM and also the key parts of Direct Marketing.
3.1. Data Mining
DM as an analytic process is designed to explore large amounts of typically business- or market related data. Consistent patterns and systematic relationships between variables are searched. DM itself is a step in the KDD (Knowledge Discovery in Databases) process. The KDD process is the process of using data mining methods, i.e. classification and clustering, to extract new knowledge from databases.27 It is the task of extracting interesting and new knowledge from large databases or similar deposits of information. Following figure28, changed for the purpose of this thesis in little detail from the original, gives an overview of the steps which comprise the KDD process. Brachmand & Anand developed this practical overview to emphasize the interactive and iterative nature of the whole KDD process.
Abbildung in dieser Leseprobe nicht enthalten
Figure 3: “ An overview of the steps comprising the KDD Process ” , Fayyad, p. 10. For original see Appendix 2.
3.1.1 Data Mining Methods
Besides clustering, classification is the most common method in Data Mining.29 That is why both methods are discussed in more detail in the two following chapters. A short overview of some widely known methods, including mentioned clustering and classification, is given here.
Classification is the process of categorizing data in order to understand and to communicate it to others. By using this type of supervised learning, objects are sorted into a predefined set of classes. Examples of classification tasks are for example assigning keywords to articles, determining which phone number corresponds to facsimile machines and assigning industry codes and job designations on the basis of free text job descriptions.30
Clustering is segmenting a heterogeneous group into subgroups, or clusters. It is also called unsupervised learning because objects are sorted into undefined categories. In difference to classification, clustering has no pre-determined groups but instead groups are made though self-similarity in the population. Elements of groups share a common set of properties.31
Estimation deals with continuously valuating outcomes.32 It is used to come up with a value for variables such as income, height and weight. Estimation can also be used to estimate the chance that someone being more or less likely to be interested in a product, this is usually signified by a number between 0 and 1.
Affinity Grouping is the method of deciding which things go together.33 This is also known as a Market Basket Analysis since the typical example is to see what goes together in a supermarket shopping cart. Store owners try to place products on the shelves according to the way the products associate, an example is cat food and kitty litter. If a person buys cat food he is also interested in kitty litter and vice versa.
3.1.1.1 Data Clustering
Data clustering is, as mentioned above, a way to divide a population into subgroups in order to simplify the location of the data required. There are several methods to accomplish this task and some of those methods are presented here below.
[...]
1 Fayyad, 1996, p.13
2 Gessaroli, 1995, p.64
3 Forcht, 1999, p. 192
4 http://www.database-marketing.de/mininghome.htm
5 http://www.database-marketing.de/mininghome.htm
6 Fayyad, 1996, p.9
7 Waiyamai, 2001
8 Waiyamai, 2001
9 Kahan, 1998, p. 491
10 The economist, 1997, p.47
11 Forcht, 1999, p.193
12 Kahan, 1998, p.491
13 Waiyamai, 2001
14 Gummesson, (2000) p. 15
15 Gummesson, (2000) p. 58
16 Gummesson, (2000), p. 59
17 Gummesson, (2000), p. 19
18 The Merriam-Webster Dictionary, (1995), p. 712
19 Gummesson, (2000), p. 179
20 Burrell & Morgan, (1992), pp. 22
21 Burrell & Morgan, (1992), pp. 3
22 Burrell & Morgan, (1992), p. 5
23 Burrell & Morgan, (1992), p. 6
24 Burrell & Morgan, (1992), pp. 23
25 Gummesson, (2000), p. 63
26 Babbie, (2000), pp. 91
27 Fayyad, (1996), p. 9
28 Fayyad, (1996), p.10
29 Barry/Linoff, (1997), p. 52
30 Fayyad, (1996), p. 13
31 Fayyad, (1996), p. 14
32 Barry/Linoff, (1997), p. 52
-
Upload your own papers! Earn money and win an iPhone X. -
Upload your own papers! Earn money and win an iPhone X. -
Upload your own papers! Earn money and win an iPhone X. -
Upload your own papers! Earn money and win an iPhone X. -
Upload your own papers! Earn money and win an iPhone X. -
Upload your own papers! Earn money and win an iPhone X. -
Upload your own papers! Earn money and win an iPhone X. -
Upload your own papers! Earn money and win an iPhone X. -
Upload your own papers! Earn money and win an iPhone X. -
Upload your own papers! Earn money and win an iPhone X. -
Upload your own papers! Earn money and win an iPhone X. -
Upload your own papers! Earn money and win an iPhone X. -
Upload your own papers! Earn money and win an iPhone X. -
Upload your own papers! Earn money and win an iPhone X.