Ch. 10
Building Knowledge-Driven DSS and Mining Data

The following five criteria should be carefully considered when evaluating vendor software for either mining data or building Knowledge-Driven DSS.

Cost. With the significant costs of technology and the rapid advancement of new technologies, companies want affordable packages. A development environment with multiple tools is often better than purchasing a more specialized development package. MIS staff want to learn software that can be applied to a wide variety of problems

Scalability. Companies need development software that will easily integrate with existing software applications and hardware platforms. Many Knowledge-Driven DSS need to be distributed to users so Web technologies are often appropriate. Some observers want more managers and analysts to have data mining tools so a distributed, scalable solution is also an issue in statistical analysis and knowledge discovery.

Security. With the increase in shared data, there is an increasing concern regarding the security of DSS knowledge and large databases. Both rule bases and behavioral data that will be mined need to be protected. Security is easily overlooked in developing knowledge applications.

Development features. Knowledge-Driven DSS are not usually standard "off-the-shelf" packages. It is important that packages allow for easy development of customized capabilities, rule input and maintenance. If uncertainties, frames or other capabilities are part of the development environment, then the package needs to help ensure that features and capabilities are used appropriately.

Ease of installation and use. Managers and MIS staff want software packages that are easy to install and require minimal training. This criterion is especially important with end-user data mining tools.