Freedom for Data Mining Professionals

Don’t limit yourself

Limitations you struggle with every day are gone. KXEN components have no practical variable limitation and eliminate the need for feature space reduction. You know that not every data set contains useful information. With KXEN you can identify these cases quickly and easily, preventing a long and expensive search for information without usable results. KXEN is so fast that you can even build models for problems you couldn’t afford to spend time on before.

Get answers faster than you thought possible

Let KXEN take care of the repetitive work. KXEN automates the data preparation process, reducing weeks of work to hours. Training a KXEN model can take minutes—or even seconds. Spend your time on the challenging problems, not on the mechanics.

Achieve both accuracy and reliability

Take advantage of KXEN’s speed to build more models—change training data, add to it, or add new variables until you are confident to have a model that is both accurate and reliable. Traditional approaches to data mining over-emphasize quality at the expense of reliability: you get models with a low training error, but they don’t produce the same results on new data. KXEN models automatically find the best compromise between fit and robustness. Even better, the KXEN information indicators tell you exactly how accurate and how reliable each model is.

Make data mining integral

While other tools strive to be standalone enterprise applications, KXEN components are designed to be embedded into your existing systems and processes. Well-documented interfaces allow easy integration into any environment. You’re free to access text files, relational databases and native sources like SAS or SPSS. User interfaces are available in wizard and expert mode. A command line interface allows easy creation of prototypes. Models can be deployed instantly without having to use a programmer.