This chapter overviews data mining starting with an explanation of the data mining methods used the most. Data mining methods are explained together with recommendations of when and how to use them and how to iteratively combine different methods. The methods are explained briefly to understand their role in projects. One of the most important topics that the analysts (readers) have to learn is how to combine different methods in the same analysis and how to use that approach to unlock the synergy effect.
Top3.1 Introduction
As shown, customer churn prediction is one of the most important problems in customer relationship management. Its major aim is to retain valuable customers helping to maximize the profit of a company. To predict whether a customer will be a churner or non-churner, there are a number of data mining techniques applied for churn prediction and variety of researches that can be found and are conducting as we are writing this book. Since enterprises in the competitive market mainly rely on the profits which come from customers, companies are trying to focus on confirmed customers that are the most fertile source of data for decision making. We can say that highly competitive organizations realize that retaining existing and valuable customers is their essential resource for surviving to survive in their industries. As shown, to create and retain customers is difficult and costly in terms of marketing. Consequently, this leads to the importance of churn management. As customer churning will likely result in the loss of businesses, churn prediction has received increasing attention in the marketing and management literature over the past time. It shows that a small change in the retention rate can result in significant impact on businesses. Customer churn can be regarded as customers who are intending to move their custom to a competing service provider. Therefore, many companies need to assess their customer’s value in order to retain or even cultivate the profit potential of the customers. In order to effectively manage customer churn for companies, it is important to build a more effective and accurate customer churn prediction model. In the literature, data mining techniques have been used to create the prediction models.
Data mining has emerged over recent years as an extremely powerful approach to extracting meaningful information from large databases and data warehouses. Since the increased computerization of business transactions, improvements in storage and processing capacities of computers, as well as significant advances in knowledge discovery algorithms, those all have contributed to the evolution of the data mining. The methodology of data mining views the discovery of information from a database as a four-step process (Tsai, Lu, 2010):
- •
Business problem must be identified,
- •
After the problem is defined and related data are collected, the next step is to process the collected data by data transformation, data cleaning, etc. for the later mining process,
- •
Apply some specific mining algorithm(s) over the processed data (like prediction/classification algorithms),
- •
Mining result is evaluated to examine whether the finding is useful for the business problem.