Java Reference
In-Depth Information
High-quality products lead to improved reputation of the organization within its
industry and help to drive sales. In addition, profitability improves through the
reduction of return materials allowances and field service calls.
2.2
Data Mining in Industries
The cross-industry solutions presented in the previous section are
readily applicable to a variety of industries. In this section, we char-
acterize some specific industry problems for which data mining can
be applied. In the discussion, we highlight some of the cross-industry
solutions as well as other problems where data mining can be
applied.
2.2.1
Financial Services
Financial services is an umbrella term that includes banking,
insurance, and capital markets. Within banking, many opportuni-
ties exist for the use of data mining, including credit scoring, credit
card fraud detection, cross-sell, as well as customer relationship
management issues including response modeling, acquisition, and
retention [SAS 2001] [SAS 2002].
In [Wu 2002], we find that
within the financial services industry, credit card issuers have been using data
mining techniques to detect potentially fraudulent credit card transactions. When a
credit transaction is executed, the transaction and all data elements describing the
transaction are analyzed using a sophisticated data mining technique called neural
networks to determine whether or not the transaction is a potentially fraudulent
charge based upon known fraudulent charges. This data mining technique yields a
predictive result. While the prediction may or may not be correct, this technique
requires the system to learn various patterns and characteristics of transactions so to
fine-tune its prediction capabilities. By utilizing data mining, credit card issuers
have decreased and mitigated losses due to fraudulent charges.
While neural networks have traditionally been used in this industry,
they also suffer from problems such as scalability and difficulty
converging on an optimal solution. Other algorithms, like Support
Vector Machine as discussed in Chapter 7, overcome these types of
problems.
Within banking, the advent of the Basel II accord, final version
issued June 2004, created a huge opportunity for data mining. Basel II is
the result of deliberations among central bankers from around the
world and the Basel Committee on Banking Supervision in Basel,
Search WWH ::
Custom Search