Information Builders Updates RStat Data Mining Tool
February 9, 2010 Alex Woodie
Information Builders last week announced the general availability of WebFOCUS RStat version 1.2, a new release of its software for building data mining and predictive analytics applications. The software doesn’t run on the System i server, but, like most Information Builder’s products, it can access data residing on that platform.
If you’ve ever watched the hit television show “Numbe3rs,” you’ve seen the powerful effect that mathematics and statistics have on our daily lives. While Charlie Eppes (played by David Krumholtz) uses his way with numbers to fight crime, people involved in all walks of life are finding new and creative uses for statistics, including launching new products, improving profitability, and even saving people’s lives.
One of the most popular statistical packages available today is R, an open source programming environment has become the de facto standard for developing statistical applications. Information Builders used an open source variant of R, called Rstat, as the basis for its implementation in WebFOCUS RStat.
Information Builders started working on WebFOCUS RStat nearly two years ago. The company initially announced the product in June 2008, and expected the first release of the product to be available by the end of the year. By the time the company launched WebFOCUS version 7.6.9 last summer, RStat was a fully recognized component of the BI suite. The company even tried using WebFOCUS RStat to predict the winner of the World Series last fall (but the Yankees won instead of the Phillies).
With last week’s launch of WebFOCUS RStat 1.2, Information Builders has fully integrated Rstat into WebFOCUS and made it a full component of the business intelligence software suite. This integration makes it easier to accomplish common tasks associated with data mining and statistical activates, such as accessing and manipulating data, which typically account for 90 percent of the cost of data mining projects, according to the company.
The closer integration of RStat and WebFOCUS also helps to reduce costs, the company says. Instead of buying and configuring separate statistical and application servers for data mining tasks, users of WebFOCUS RStat can use the same underlying hardware that WebFOCUS was already utilizing (with the one exception being that RStat does not run on the System i server, whereas the overall WebFOCUS suite can be deployed on that platform, so this advantage is lost to System i customers.)
Version 1.2 adds several other new features, including new “survival analysis” statistical modeling and scoring routines that use Kaplan-Mayer and COX regression techniques. These techniques will be useful for helping users to predict the outcomes of events in pharmaceutical, healthcare, and government settings, Info Builders says. New charts have also been added for survival analysis.
The library of scoring routines in WebFOCUS RStat has been expanded to support advanced models such as neural networks, which are used to solve artificial intelligence problems. This release also brings new testing capabilities that support data mining as well as standard statistical analysis, the vendor says.