Written by: STATISTICA News 3/25/2008 9:19 AM
TULSA, OK (March 25, 2008) - Dr. Bob Nisbet, a recognized expert on data mining and the author of many important publications in this area, has recently performed a quick test of STATISTICA Data Miner with built-in Data Miner Recipes. Dr. Nisbet stated that he “liked the way the new Recipes interface leads the user through common data mining tasks” and he particularly liked the redundancy check.
The evaluation copy that Dr. Nisbet was provided did not include documentation, which led to an unintended “test of intuitiveness” of that new UI – Dr. Nisbet was still able to use the Recipes facility. Moreover, using the default parameters in STATISTICA Data Miner, KXEN (version 3.3.3) and SPSS Clementine Neural Net (Version 11.0), he found that the automated function of STATISTICA Data Miner Recipes was as automatic (or nearly so) as these others tools and yet provided superior performance. Dr. Nisbet’s sample data set was a telecommunication data set with 10,000 rows and the tests were conducted using out-of-the-box parameters and configurations for all three software packages. Overall performance was measured on the basis of Global Accuracy as measured by (True-Positives + True-Negatives)/N.