Bayer Links


Winner - Hermes Award 2007

Data Mining Technologies

Gaining process understanding from plant data

Click on the image to magnify.

Our technology platform

K.wiz (thinkAnalytics Ltd.) is the software platform for our data mining applications. K.wiz is a state-of-the-art environment that allows the interactive specification and manipulation of complex data mining workflows. These analysis workflows can be easily embedded in other software environments like web-based applications.

In data mining for process data analysis, highly specialized methods are required that are not found in standard commercial data mining packages. A prominent example is the ability to adequately analyze data from dynamic systems.

Based on our many years of experience as a leading technology provider, we have developed a data mining expert toolbox tailored to the needs of process analysis and containing a comprehensive set of currently more than 100 methods for data preprocessing, analysis and validation:

  • decision tree analysis;
  • subgroup discovery;
  • cluster analysis;
  • rule and attribute statistics.

All these methods can be used from our standard K.wiz platform.

Data mining can only reveal its full potential if a validated link to the process data can be created and if a reliable interpretation of the results can be ensured. Bayer Technology Services is able to provide both of the above as its data mining technology is integrated into a large pool of mathematical technologies. These range from special know-how in accessing process data archives through to expertise in and tools for rigorous modeling, neural networks and hybrid modeling (a combination of the two). Modeling technologies are particularly important as only they enable quantification of the data mining results and thus their application in process optimization.