SAS has designed Enterprise Miner 4.1 to cull the gems from a vast assortment of enterprise data by unearthing trends, explaining known outcomes, predicting future results, and identifying factors that will ensure a desired effect.
When combined with SAS data warehousing and OLAP (online analytical processing) technologies, Enterprise Miner addresses the full spectrum of knowledge discovery, SAS says. Although Enterprise Miner provides a range of integrated models and algorithms, its graphical user interface and automated framework are set up so that those with little statistical expertise can easily use these tools.
The automated data-mining process helps users model new questions that they might not have thought to ask and to refine data at each stage of discovery. “A sophisticated set of tools allows marketing analysts to fine-tune the process,” said Wayne Thompson, worldwide data-mining program manager at SAS Institute.
Data, Data Everywhere
The need for data mining continues to grow. “Over the last decade, the proliferation of data has been massive,” Thompson told CRM Buyer Magazine, “particularly in such industries as telecommunications, financial services and pharmaceuticals. Without data mining to make inferences into the future, these organizations are at a loss.”
Tough economic times have made this even more of a challenge. “As the economy has gone south, there has been a huge pressure on CIOs to justify the expense of what they bought,” Forrester analyst Nate Root told CRM Buyer, “and how to dig data — such as the customer’s historical behavior — from what they bought.”
Enterprise Miner’s predictive modeling techniques help identify the user’s most profitable customers, determine the products most likely to be purchased, detect fraudulent behavior and help increase response rates from direct-mail, telephone, e-mail and Internet-delivered campaigns. The technology also helps improve customer profitability through more accurate credit scoring and saves on downtime by applying predictive maintenance to manufacturing sites.
Added in June, Text Miner 8.2 expands data-mining capabilities to text-based documents — for example, from e-mail, call center records and warranty claims. Text-processing and analysis tools enable the user to uncover underlying themes, automatically group documents into topical clusters, and classify documents into predefined clusters.
Combining Enterprise Miner and Text Miner provides better predictive models of what customers will purchase in the future, Thompson said. “We’ll continue to work on the integration of text mining and data mining.”
In the next year or two, SAS plans to improve the performance of its mining algorithms to allow the distribution of work over multiple CPUs and to develop new variable-selection algorithms that handle high-dimensional data. In addition, Thompson said, the company aims to make data-mining tools available to more people within large enterprise organizations.
Also ongoing for SAS — as well as other data-mining companies, including Angoss, IBM, KXEN and SPSS — is participation in the Data Mining Group, a consortium that is forging data-mining standards, such as PMML, or predictive model markup language. That XML-based language provides a way to define statistical and data-mining models and to share them among PMML-compliant applications.