script src="//" data-sumo-site-id="aa6c3d60b32e5862815eb98663e1413c4f7864fbc2afb58dd5f1de51d4a7cfe4/script Talkpropeller's Articles: 2013-04-07

Art for art sake

Art for art sake

Friday, April 12, 2013

The beginning of the universe, for beginners - Tom Whyntie

Contemporary Leadership Theories:

In describing  the major leadership theories, it has been reasonable to start with the Great Man Theory (GMT) as it was the first leadership theory (Edgar F 1954). Therefore, it understandable that some of the current leadership literature has its roots and originality from GMT(Helen L. Eckmann, 2008 ). 
·                     Great Man Theory
The great man theory is considered the oldest leadership theory (Edgar F. Borgatta, Robert F. Bales and Arthur S. Couch, 1954). This theory was popularized in 1840 by Scottish writer Thomas Carlyle. Carlyle viewed leadership by the impact of “great men” or “heroes” whom have strong influential on their societies and individuals in a way that leaded to a significant historical impact, usually an impact that had been highly desired by the nation. 
The great men theory is mostly associated with historical leaders in historical events; those leaders uniquely had their name printed in history for lifetime. Such great leaders are: Alexander the Great, Napoleon, Julius Caesar, Queen Elizabeth I, Abraham Lincoln etc. In addition, there are also some contemporary leaders such as H.H. S. Zayed Bin Sultan, King Faisal, and Mahatma Gandhi. (
Major attributes with Great Men theory are:
·                     The leaders are born and not made
·                     Great leaders can arise in extreme need
The theory had focused on male gender as great leader, and has totally ignored the other gender. In addition, early research on great men theory was based on legendary great men who had lead their nation through historical events, most of these leaders were from the aristocracy, and few from lower classes.
This theory had been causing dilemma and controversial issue with many leaders and researchers, some accepted this theory in certain specific situations, others believe this theory is highly depending on individual and discriminate group work, more interestingly, this theory discourage the group empowerment and depends mostly on decisive leader who can make his group listen, obey, and follow him.( Abdull, 2009).

·                      Trait Theory:
The trait theory arose from the previous first theory the Great Man theory. “The trait approach was one of the first systematic attempts to study leadership”, according to Peter Northouse (2004). The traits theory describes leaders by those who have identifying key characteristics of traits (Bolden, R., Gosling, J., Marturano, A. and Dennison, P, 2003). It basically emphasize that leaders are born with inherited traits, in which these specified traits make good leaders. In addition, good leaders should have balance combination of “leadership Traits (Mark. S, 2010; anonymous, 2008).
The early research on traits theory was based on studying leadership’s characteristics that are common between successful leaders. This study was done through psychological focus of the day of leaders (Eryn Travis , 2011).
One of the largest reviews on Traits theory was accomplished by Dr. Ralph Melvin Stogdill (David Boje, 2008). Stogdill had reviewed over 287 traits studies between 1904 to 1970, the first trait study was 1904- 1948, the following one was 1949 to 1970 (David Boje, 2008;  David, 2004). Stogdill said: “a person does not become a leader by virtue of the possession of some combination of traits” (1948). His research outputs illustrated that leaders are distinguished from other people by the followings: Adaptable to situations, alert to social environment, ambitious and achievement-oriented, assertive, cooperative, decisive, dependable, dominant (desire to influence others), energetic (high activity level), persistent, self-Confident, tolerant of stress, and responsibility.
Stogdill had included his list of traits another list of skills that are as necessary as  list of traits. These main skills are: clever (intelligent), conceptually skilled , creative, diplomatic and tactful, fluent in speaking , knowledgeable about group task, organized (administrative ability), persuasive, and socially skilled.

Mann (1959) had also reviewed over 1,400 findings of the relationship of personality and leadership in small groups. His findings were clear and supports in somewhat Stogdill findings; they both suggested that such traits are not reliable predictors of who will emerge into leadership roles (Mann 1959; Stogdill 1948). Mann found in his studies that leaders traits are: intelligence, masculinity, dominance, adjustment, extroversion, and conservatism.

In 1986 a group of researcher led by Robert G. Lord, reviewed Mann’s observations of traits, and their conclusion was that the first three characteristics (1. intelligence, 2. masculinity, 3. Dominance) are related to followers how they perceives leaders.

In conclusion, it is clear that researchers on the filed of traits theory failed to develop a definitive list of leadership traits. In addition, traits studies had failed in studying leaders traits in during the impact of situations. Trait theory had not adequately linked leaders traits with the important outcomes which are team performance nor task achievement(David, 2004). However, the trait theory still somehow a validate study, as it a result of century of study, yet it provides a benchmark tool for any person who interested to evaluate their personal leadership attributes (David, 2004). 
Behavioral Theory:
Historically, after years of dominating that  great leaders are only born, then domination  that leaders born with certain traits, the behavior theory started to take place in the late 1950s early 1960s ( Leow Fook Thye, 2010).
In the opposite to the old assumptions of the previous theories of leadership that leaders are born (Thomas Carlyle, 1840; , )the behavior theory claim that leaders can be learned and trained. The behavior theory is a leap of traits theory. The behavior theory believes personal’s leadership potentials can be trained into effective leadership competencies, moreover leadership should be merge with management to achieve it maximums potential outcomes (Thye,2010; Bass, 1990). The behavior theory focuses on what leader “does”, it basically about adopting common leadership behaviors and “do” it (Penn,---). Similar, to what extent that Thy and Bass concluded that leadership can be trained, Kirkpatrick and Locke (1991) had both contend that even though some leaders born with leadership’s traits, other individuals can learn leadership traits.

What is leadership

Leadership is the ability of a superior to influence the behavior of a subordinate or group and persuade them to follow a particular course of action

Defining the meaning of leadership in business world has been driven from different meanings, and yet there is no consensus definition of leadership. These meanings are basically different as they are driven from:
·                     conclusion of human philosophical perspectives
·                     a result of several situations that people were involved in and believed that what they have done or achieved is the right “leadership” definition
The knowledge of leadership has been studied over 80 years, and yet has still not become with unequivocal or definitive conclusion about it core essence.  There are more than 850 definitions of what constitutes leadership (Warren Bennis,  2003).
The common definition of leadership is the ability of a person to lead people to achieve objectives. (Massie) Leadership is also influenced by individuals prospective, many leaders vision leadership as being superlative among his/ her followers and influence them and persuade them to follow a particular course of action. (Chester Bernard).   
It is extremely surprisingly how difficult to define leadership (RAUL A. ANGULO, 2001; Joel Broughton 2010).  The word leadership has been greatly affected by the political leadership influence where the word “lead” is still being used and associated with politics and power of control.
Leadership definitions have been also classified based on how people perceive them. There is huge view leadership as series of traits or characteristics, other as skills and knowledge, and some believe leadership is a process (Steve Wolinski 2010). 
·                     Leadership as a process:  it places importance on social interaction and relationship. It defines leadership as a process of giving purpose to collective effort (Jacobs & Haques), in addition influencing the activities of an individual or group toward goal achievement in a given situation (Hersey & Blanchard; Rauch & Behling). This process also acquire influencing the activities of followers through social interaction such as communication (Donelly), interpersonal influence that has been exercised in a  directed situation toward the attainment a specified goal (Tannenbaum, et al). 
·                     Leadership as traits and skills is being widely perceived by many researchers as knowing yourself as leader(Bennis), having the vision, excellent communication skills, embodying your followers by building a trust, getting emotionally evolved, while establishing a creative working environment(Bennis, Richards and Engle, Ciulla). It also complex moral relationship between people, based on obligation and commitment for shared good (Ciulla).
·                     Leadership also has been classified as part of art and part of (perhaps even “immature”) science (Philip, 2011). However, this concept of leadership “art and science” is still causing large controversial discussions with many leadership experts. The dilemma is many argue that leadership is only art, and will never be a science as science concludes with objective, and leadership never ends with objective (Maxwell, J.C, 2011) as it is entirely subjective, therefore leadership is an art. On the other side, those who argue leadership is a large science insists that this science comes from different prospective linked directly and indirectly to leader’s moral abilities and knowledge; such as:  processes and systems developed driven from the examination of how we think, psychological study (Gwyn Teatro, 2010).

Data mining concept and techniques used in CRM



  1. Overview

Data Mining or DM has enjoyed its popularity in recent years, with advances in both research and commercialization. Data mining is an iterative process within which progress is defined by discovery, through either automatic or manual methods. Data mining is most useful in an exploratory analysis scenario in which there are no predetermined notions about what will constitute an "interesting" outcome. Data mining is the search for new, valuable, and nontrivial information in large volumes of data. It is a cooperative effort of humans and computers.Its first generation of data mining research and development has yielded several commercially available systems. Its research has tended to be fragmented. They are divided into several distinct pockets without a comprehensive framework. Many researchers have continued to work largely within the parameters of their parent disciplines, building upon existing and distinct research methodologies. Even for a common problem like how to cluster a dataset researchers apply different techniques, different perspectives on what the important issues are, and different evaluation criteria. While different approaches can be complementary, and diversity is ultimately a strength, better communication across disciplines is required if DM  is to forge a distinct identity with a core set of principles, perspectives, and challenges that differentiate it from each of the parent disciplines. Further, while the amount and complexity of data continues to grow rapidly, and the task of distilling useful insight continues to be central, serious concerns have emerged about social implications of data mining. Addressing these concerns will require advances in our theoretical understanding of the principles that underlie Data Mining algorithms, as well as an integrated approach to security and privacy in all phases of data management and analysis.

As for the concept of DM, the word "process" is very important here. Even in some professional environments there is a belief that data mining simply consists of picking and applying a computer-based tool to match the presented problem and automatically obtaining a solution. This is a misconception based on an artificial idealization of the world. There are several reasons why this is incorrect. Data mining is not simply a collection of isolated tools, each completely different from the other, and waiting to be matched to the problem. Instead it lies in the notion of matching a problem to a technique. Only very rarely is a research question stated sufficiently precisely that a single and simple application of the method will suffice. In fact, what happens in practice is that data mining becomes an iterative process (as defined earlier). Here, one studies the data, examines it using some analytic technique, decides to look at it another way, perhaps modifying it, and then goes back to the beginning and applies another data-analysis tool, reaching either better or different results. This can occur many times; each technique is used to probe slightly different aspects of data—to ask a slightly different question of the data. Essentially, what is being described here is a voyage of discovery that makes modern data mining exciting. Still, data mining is not a random walk through the space of analytic techniques but a carefully planned and considered process of deciding what will be most useful, promising, and revealing.

Having considered the DM aspect, it is also important to consider its application. Here, Customer Relationship Management (CRM) is chosen. It emerged in the last decade to reflect the central role of the customer for the strategic positioning of a company. CRM takes a holistic view over customers. It encompasses all measures for understanding the customers and for exploiting this knowledge to design and implement marketing activities, align production and coordinate the supply-chain. CRM puts emphasis on the coordination of such measures, also implying the integration of customer-related data, meta-data and knowledge and the centralized planning and evaluation of measures to increase customer lifetime value. CRM gains in importance for companies that serve multiple groups of customers and exploit different interaction channels for them. This is due to the fact that information about the customers, which can be acquired for each group and across any channel, should be integrated with existing knowledge and exploited in a coordinated fashion.

It should be noted, however, that CRM is a broadly used term. It covers a wide variety of functions, not all of which require data mining. For example, CRM as a product/market segment was first introduced by companies such as Siebel and Oracle, with many other players, including SAP, PeopleSoft and Microsoft joining subsequently. The initial set of products mostly support easy management of information for customer facing functions, including contact management, sales force automation, etc. Applying data mining to better understand customers, and its use for relationship management, constitutes a more recent phenomenon. These functions include marketing automation (e.g., campaign. management, cross- and up-sell, customer segmentation, customer retention), sales force automation (e.g., contact management, lead generation, sales analytics, generation of quotes, product configuration), and contact center management (e.g., call management, integration of multiple contact channels, problem escalation and resolution, metrics and monitoring, logging interactions and auditing), among others. We focus on how backend data mining and analytics can make these functions more effective.

The role of CRM is in supporting customer-related strategic measures. Customer understanding is the core of CRM. It is the basis for maximizing customer lifetime value, which in turn encompasses customer segmentation and actions to maximize customer conversion, retention, loyalty and profitability. Proper customer understanding and actionability lead to increased customer lifetime value. Incorrect customer understanding can lead to hazardous actions. Similarly, unfocused actions, such as unbounded attempts to access or retain all customers, can lead to decrease of customer lifetime value (law of diminishing return). Hence, emphasis should be put on correct customer understanding and concerted actions derived from it.

The reality of CRM, especially in large companies, looks quite different from the central coordination and integration
• Information about customers flows into the company from many channels, but not all of them are intended for the acquisition of customer-related knowledge.

• Information about customers is actively gathered to support well-planed customer-related actions, such as marketing campaigns and the launching of new products. The knowledge acquired as the result of these actions is not always juxtaposed to the original assumptions, often because the action-taking organizational unit is different from the information-gathering unit. In many cases, neither the original information, nor the derived knowledge is made available outside the borders of the organizational unit(s) involved. Sometimes, not even their existence is known.

• The limited availability of customer-related information and knowledge has several causes .Political reasons, e.g. rivalry among organization units, are known to lead often in data and knowledge hoarding. A frequently expressed concern of data owners is that data, especially in aggregated form, cannot be interpreted properly without an advanced understanding of the collection and aggregation process. Finally, confidentiality constraints, privacy considerations and law restrictions often disallow the transfer of data and derived patterns among departments.

• In general, one must assume that data gathered by an organization unit for a given purpose cannot be exported unconditionally to other units or used for other purpose and that in many cases such an export or usage is not permitted at all.

• Hence, it is not feasible to strive for a solution that integrates all customer-related data into a corporate warehouse. The focus should rather be in mining non-integrated, distributed data while preserving privacy and confidentiality constraints

2.         Literature review

As for the background literature covered for this study, they include the book by I.H. Witten and E. Frank, Data Mining: Practical Machine Learning Tools and Techniques with Java Implementations, Morgan Kaufmann, San Francisco, 2000. This covers the learning tools and techniques in data mining. As such, one should have some knowledge of the concepts and terminology associated with database systems, statistics, and machine learning. As for the data analysis per se one considers this book by M. Berthold and D.J. Hand, Intelligent Data Analysis: An Introduction, Springer, Berlin, 1999, as the ideal one. It’s very useful for understanding data mining and its application. However, for one wanting to cover the statistical analysis in data mining, it is useful to use the book by T. Hastie, R. Tibshirani and J. Friedman, The Elements of Statistical Learning: Data Mining, Inference and Prediction, Springer, New York, 2001. This is an important literature for statistical study and its application. As most of data mining involves some form of statistical review, this is the most appropriate source to consider. In most study, one needs to understand the principles of how it works, thus for the principles of data mining it is necessary to cover this book by D. Hand, H. Mannila and P. Smyth, Principles of Data Mining, MIT Press,2001. As to the subject of data mining itself: what is it? One may find that it is relevant to consider T.M. Mitchell, Machine Learning, McGraw- Hill, New York, 1997...

3.         Data mining tasks and techniques used
Data mining commonly involves four classes of tasks
·         Clustering - is the task of discovering groups and structures in the data that are in some way or another "similar", without using known structures in the data.
·         Classification - is the task of generalizing known structure to apply to new data. For example, an email program might attempt to classify an email as legitimate or spam. Common algorithms include decision tree learning, nearest neighbor, naive Bayesian classification, neural networks and support vector machines.
·         Regression - Attempts to find a function which models the data with the least error.
Association rule learning - Searches for relationships between variables. For example a supermarket might gather data on customer purchasing habits. Using association rule learning, the supermarket can determine which products are frequently bought together and use this information for marketing purposes. This is sometimes referred to as market basket analysis.

Three are two sections of DM techniques can one can find. They have been broken up based on when the data mining technique was developed and when it became technically mature enough to be used for business, especially for aiding in the optimization of customer relationship management systems.  Thus this section contains descriptions of techniques that have classically been used for decades the next section represents techniques that have only been widely used since the early 1980s. By doing so, it is able to help the user to understand the rough differences in the techniques and at least enough information to be dangerous and well armed enough to not be baffled by the vendors of different data mining tools.

The main techniques that we will discuss here are the ones that are used 99.9% of the time on existing business problems.  There are certainly many other ones as well as proprietary techniques from particular vendors - but in general the industry is converging to those techniques that work consistently and are understandable and explainable. The other commonly used techniques are statistical. By strict definition statistical techniques are not data mining.  They were being used long before the term data mining was coined to apply to business applications.  However, statistical techniques are driven by the data and are used to discover patterns and build predictive models.  And from the users perspective one will be faced with a conscious choice when solving a "data mining" problem as to whether one wish to attack it with statistical methods or other data mining techniques. 


From the above study one is able to identify key data mining challenges and opportunities of which one can point out some conclusions.

It can be said that Non-trivial results almost always need a combination of DM techniques. Chaining/composition of DM, and more generally data analysis, operations is important. In order to analyze CRM data, one needs to explore the data from different angles and look at its different aspects. This should require application of different types of DM techniques and their application to different “slices” of data in an interactive and iterative fashion. Hence, the need to use various DM operators and combine (chain) them into a single “exploration plan”.

Apart from that there is a strong requirement for data integration before data mining. In both cases, data comes from multiple sources. For example in CRM, data needed may come from different departments of an organization. Since many interesting patterns span multiple data sources, there is a need to integrate these data before an actual data mining exploration can start.

Often, one encounters diverse data types which require the integrated mining of diverse and heterogeneous data. In CRM, while dealing with this issue is not critical, it is nonetheless important. Customer data comes in the form of structured records of different data types (e.g., demographic data), temporal data (e.g., weblogs), text (e.g., emails, consumer reviews, blogs and chat-room data), (sometimes) audio (e.g., recorded phone conversations of service reps with customers).

It is also the issue of privacy and confidentiality for data and analysis results they are a major issue. In CRM, lots of demographic data is highly confidential, as are email and phone logs. Concern about inference capabilities makes other forms of data sensitive as well—e.g., someone can recover personally identifiable information (PII) from web logs.

Most significant of all, is the legal considerations influence what data is available for mining and what actions are permissible. In some countries it is not allowed to combine data from different sources or to use it for purposes different from those for which they have been collected. For instance, it may be allowed to use an external rating about credit worthiness of a customer for credit risk evaluation but not for other purposes. Ownership of data can be unclear, depending on the details of how and why it was collected, and whether the collecting organization changes hands.

One can also see that In CRM, as in many DM applications, discovered patterns are often treated as hypotheses that need to be tested on new data using rigorous statistical tests for the actual acceptance of the results. This is even more so for taking or recommending actions, especially in such high-risk applications as in the financial and medical domains. Example: recommending investments to customers (it is actually illegal in the US to let software give investment advice).

It is also important that once data mining has been conducted with promising results, how to use them in the daily performance task is critical and it requires significant research effort. It is common that after some data results are obtained, the domain users do not know how to use them in their daily work. This research may require the participation of business and marketing researchers. Another way to accommodate actioning mechanisms is to integrate them into the knowledge discovery process by focusing on the discoveries of actionable patterns in customer data. This would make easier for the marketers or other domain experts to determine which actions should be taken once the customer patterns are discovered.

Incorporating prior knowledge has always been a problem in practice. Data mining tends to find many pieces of patterns that are already known or redundant. Incorporating prior domain knowledge can help to solve these problems, and also to discover something novel. However, the difficulties of incorporating domain knowledge result in little progress in the past. There are a number of reasons for this. First of all, knowledge acquisition from domain experts is very hard. This is well documented in AI research, especially in the literature of expert systems building. Domain experts may know a lot but are unable to tell. Also, many times, domain experts are not sure what the relevant domain knowledge is, which can be very wide, although the data mining application itself is very narrow. Only after domain experts have seen some discovered patterns then they remember some domain knowledge. The second reason is the algorithmic issue. Many existing methods have difficulty to incorporate sophisticated domain knowledge in the mining algorithm. Also, once the new patterns are discovered, it is important to develop methods that integrate the newly discovered knowledge with the previous knowledge thus enhancing the overall knowledge base. Although there is some general work on knowledge enhancement, much more needs to be done to advance this area and adapt it to CRM problems. Also, integration of these methods with existing and novel Knowledge Management approaches constitutes a fruitful area of research.