University of Oulu
INTELLIGENT SYSTEMS GROUP

Data Mining

From plain data to useful knowledge

Data mining is the process of extracting usable knowledge from data sets. The measurement data can originate from various sources, for example from a production facility, human being or a consumer electronics device. The common factor for all these sources is that they contain useful information which can be extracted from the raw measurement data with the methods we are developing.

In the process of forming the knowledge, multidisciplinary research methods are used; we advance the development, integration and application of research methods commonly used for example in statistics, pattern recognition, software and database research.

Applying research results in real world applications

The measurement data originates from applications and therefore our research challenges are application-driven. For us, the applications offer concrete cases where we can prove the effectiveness of the methods we develope. Our research results have proven to be capable of improving and enhancing the applications by providing:

  • better understanding of the application based on real measurements,
  • features only available by applying data mining methods on the measurements (e.g. prediction, recognizing the state of the application, etc.) and
  • new tools for on-line implementation of the results.

Currently we are working with applications from production, metal and mobile industries. However, our approach is not limited to any one field in particular and our research is always made in close co-operation with an application expert.

Advancing the transition of know-how into business

We have a solid track of industrial co-operation since the foundation of the group. At the moment, most of our research projects and applications originate from national and international industries. The co-operation is based on the needs and requirements arising from the industry. Based on these needs we transfer the latest, suitable knowledge in the field for the benefit of our partners.

See our Publications

See our  XPRESS Project

See our People

Production industry

Our advanced method for identifying spot welding processes can be applied to shorten the start-up times of joining processes in the automotive and electrical industries.
Mobile (ubiquitous) electronics

Research on context aware mobile electronics. We are developing methods for automatic recognition of human activities for example in every day environments.
Metal industry 

A novel approach developed for planning the properties of steel plates. Initialization of our method has significantly improved the manufacturing process of steel products.

 

Last modified: 2009-02-12