DATA MINING Desktop Survival Guide by Graham Williams
Desktop Survival Project Home Preface List of Figures List of Tables Enterprise Data Mining Process Rattle Data Explore R Data Graphics in R Understanding Data Preparing Data Building Models Evaluating Models Algorithms Apriori Bagging Bayes Classifier Boosting Conditional Trees Hierarchical Clustering K-Means K-Nearest Neighbours Text Mining Open Products Borgelt Data Mining Suite R Rattle Weka Closed Products C4.5 Clementine Enterprise Miner Equbits Foresight GhostMiner InductionEngine ODM Statistica Data Miner TreeNet Virtual Predict Appendicies Glossary Bibliography Index

Business Problems

Data mining consists of multiple data analysis and model building techniques that can be used to solve different types of problems in business. Although it is not the only solution to these problems, data mining is widely used because it suits best for the current data environments in enterprises. That is, data volume is getting bigger, data content is getting more complex, and business environment is changing faster than before. Data mining can deal with large complex data environment and provide quick solutions to catch up with the economic changes.

Below is a list of typical business problems data mining is used to solve:

Customer profiling: Building customer profiles is a necessary step in marketing, customer service and customer relationship management. Customers are not same. Knowing their profiles, we can apply different strategies to different customers. Customer profiles can be divided into different types, such as demographic profile, behavior profile and hobby profile. Data mining techniques can be used to create customer profiles from customer data. For example, the decision tree technique can be used to create demographic profiles of high spending customers from customer demographic data and recent spending data.

Customer segmentation: Customer segmentation is to divide customers into different groups in which each has a different profile and characteristics. The simplest segmentation method is to divide customers into different age groups, sex groups and income groups. Although such segmentation is useful for certain purposes and still in use in many cases, it is too coarse and no longer satisfies the new business requirements in direct marketing, recommendation and personalized service. Data mining provides techniques to create fine customer segments to satisfy the new business requirements.

Direct marketing: Direct marketing is an important means to sell products and services to existing customers. An important measure of success on a direct marketing campaign is the customer response rate to the product offer. Data mining technology can help select those targeted customers who have a high likelihood to respond the offer. The data mining involvement in direct marketing includes building models from previous campaign data and using the models to select targeted customers for new campaigns. Because the models are learnt from the previous campaign results, they can identify the targeted customers who are likely to accept the offer. This method not only increase customer response rate for the campaign, but also reduces the campaign costs.

Cross-selling: If a bank sells mobile services to bank's credit card customers or a mobile operator sells credit cards to its customers, this sales activity is called cross-selling. Cross-selling is a win-win business to the participating partners. By merging customer data from two different companies, we can use data mining to discover cross-selling opportunities and build cross-selling models to target right customers for cross-selling in two organizations.

Customer retention: In a competitive market, retaining profitable customers is a big challenge to every company. In some industry, for instance, mobile services, customer annual attrition rate can be as high as 40% in some countries. Data mining technology combined with some marketing initiatives can be used to reduce customer attrition rate. From customer behavior data, for example, service usage and charging fee data, we can use data mining techniques to beuild customer retention models that can predict which customers are likely to defect and when. Having identified these customers, we can use data mining techniques to descover characteristics of the customers. Marketing initiatives can be designed to retain these customers by offering them some incentives.

Fraud detection: Fraud is a business plague. It causes great financial damages to companies. Fraudulent activities are usually recorded in business transactions which are in large volume. Business fraud has many different forms and usually happen rarely and irregularly. Therefore, many business frauds are difficult to detect. Data mining provides useful techniques to sift out frauds from large business data.

Support further development through the purchase of the PDF version of the book.
Brought to you by Togaware.