The sum is often greater than its parts, and in data mining this becomes quite obvious when multiple technology tools are combined in a suite of paradigms. The following section showcases this new breed of multiple paradigm products.
These tools can combine features and technologies and are capable of linking analyses and their results in formats that can segment your website visitors by their demographic or lifestyle features, then present them in the form of SQL syntax, C code, or a 3D ZIP code map. Incorporated in these new tools are multiple technologies including statistics, neural networks, genetic algorithms, decision trees, and visualization modules enabling users to view relationships, extract patterns, and construct predictive models from their data.
Using one or two of these high-end tools you can do visualization, segmentation, association, classification, prediction, clustering, and optimization. Some are capable of running on scalable servers and take full advantage of their multiprocessor architectures. These data mining vendors recognize that no single technology can answer all the business questions a data warehouse or a large electronic commerce site may have about their data and as such, their products are designed to provide multiple solutions. Specifically built to deliver business results, most of these tools are able to generate C code, SQL syntax, three-dimensional graphs and maps for presentations and exporting into production systems. Some have the ability to base their analysis on millions of observations.
Tool
Vendor
Platform
Technologies
Clementine
SPSS
NT/UNIX
Back-propagation, SOM, C5.0
DataEngine
Copyright Butterworth-Heinemann 1999 under license agreement with Books24x7