New Ncr Data Mining Software Enables Businesses To More Quickly Analyze Transaction Data - TeraMiner Stats - Product Announcement

EDP Weekly's IT Monitor, July 5, 1999

To help businesses more quickly and accurately harvest the growing volumes of data from daily transactions to better understand and serve their customers, NCR Corp. last week unveiled its TeraMiner Stats data mining software, part of an NCR-developed suite of tools designed to improve data mining processes for Teradata Warehouse customers.

NCR says TeraMiner Stats overcomes the limitations of most data mining tools, which are not designed to handle the large and growing data volumes now common to most companies. According to industry analysts at International Data Corp., the amount of data in the average data warehouse will quadruple over the next three years. Rather than using a sample of data to perform the data mining analysis, TeraMiner Stats can perform statistical analysis and transformations on any or all data in the warehouse -- even if it's multiple terabytes.

With TeraMiner, NCR has moved core data mining capabilities into its Teradata database. The increased functionality in Teradata gives organizations that want to mine growing volumes of data the necessary flexibility to do so in a high-performance environment. TeraMiner Stats includes scalable pre-processing components that address critical areas of data mining, including data description, derivation, reduction, reorganization and sampling/partitioning.

By leveraging the inherent parallelism and processing power of the Teradata engine, NCR says TeraMiner Stats enables customers to use all their detail level data in performing complex data mining analyses, such as customer segmentation, propensity to buy, sales forecasting, attrition and retention.

The result is that Teradata Warehouse customers can increase their return on investment (ROI) by developing more accurate predictive models in less time. This allows customers to execute more frequent and highly tuned campaigns.

According to Forrester Research, major database players are moving data mining functionality into their database engines, enabling business- centered data mining. Forrester believes it is critical for data mining technology to move from its experts-only orientation to being an accessible, integrated technology that can be used across the enterprise.

"TeraMiner Stats streamlines data preparation and descriptive analysis -- which typically consumes between 60 and 80 percent of a company's data mining efforts and resources before they even analyze their data," says Vickie Farrell, assistant vice president of Teradata marketing for NCR. "This approach makes mining large volumes of data feasible. With TeraMiner, NCR has already delivered what other vendors are only taking about."

TeraMiner leverages Teradata's ability to handle a large number of multiple concurrent users and provides a single repository of analytic data sets, as opposed to redundant data sets across the enterprise. This enables businesses to streamline their data mining process by minimizing the number of steps required to analyze and manage that data.

Additional TeraMiner products will be released in 2000, says Farrell, providing statistical and machine-learning algorithms, as well as model deployment, monitoring and lifecycle maintenance.