Java Reference
In-Depth Information
ABCBank has an internal data mining lab (DM Lab) that is
responsible for finding data mining solutions for given business
problems. DM Lab has business analysts and data analysts that work
together to define the mining attributes and data needed for the busi-
ness problems and to define the model building process. DM Lab
coordinates with the ABCBank information technology (IT) depart-
ment to acquire the relevant data from various data sources for min-
ing. DM Lab then communicates back to IT with the final results
produced using the data. IT integrates the data mining results with
the appropriate applications, so that company executives can view
the results. Figure 11-4 depicts the high-level view of the current
scheme of integrating data mining results with the production
applications.
The current scheme is not efficient due to delays in delivering
the data mining results to the organization's information consum-
ers. Even though DM Lab generates accurate models, ABCBank is
dissatisfied with the time it takes DM Lab to produce results and
IT to incorporate those results into applications and reports. For
example, for the customer attrition problem, the
attrition model
pro-
duces good results at DM Lab. However, by the time the model is
deployed to the call center, some valuable customers have already
left. In addition, ABCBank would like to expand the use of data
Enterprise Applications
Provide DM results to specialized
analytics applications that are not
integrated with production applications
Data Mining Lab
IT Department
Specify needed attributes to IT
Collect the DM Lab specified data from
the various data sources
Acquire data from IT
Build and evaluate models
Produce deployable mining results
Monitor performance of deployed results
Provide the data to DM Lab
Integrate data mining results given by
DM Lab with the associated applications
Figure 11-4
ABCBank's current data mining model deployment process without
using JDMWS.
Search WWH ::
Custom Search