WP5 – Data Mining and Knowledge Discovery


The main objective of this work package is to build the EURECA data mining and knowledge discovery services to support the identified user scenarios with a data mining focus. These tools and services will be integrated into an advanced framework addressing both clinical research and care (also supporting this way WP6: Application, semantic reasoning and decision support).

Tools enabling the discovery of new, clinically relevant knowledge will be implemented, as well as tools for the automatic analysis of biomedical data in the form of data mining and statistical procedures.

Next to the integration of existing tools, research activities will concentrate on machine learning algorithms for the discovery of long-term patterns of low frequency. This will allow the discovery of patient safety issues that are not frequent enough to be found within the scope of clinical trials.

This Work Package will also investigate the automatic generation of research hypotheses by statistical pattern recognition. The development of new approaches for similarity learning on partly structured domains will support the researcher with help for the selection of data sets for meta analysis, and support of the transfer of analytic solutions to new trials and datasets.


5.1 - Requirements analysis and knowledge discovery scenarios

5.2 - State of the art review of existing methods and tools for hypothesis generation and association study

5.3 - Initial prototype of the generic knowledge discovery framework

5.4 - Initial services for hypothesis generation and association studies for safety risks and new research

5.5 - Refined generic knowledge discovery framework an services based on evaluation with users

5.6 - Validation of the knowledge discovery services and framework


To see D5.1 click here:

http://share.ecancer.org/EURECA/Deliverables/EURECA-D5 1